Facing a media spasm of confusion, misinformation and opportunism, representatives of the HIV research and activist communities desperately tried to correct the facts, fearing an irrational backlash against a simple and affordable treatment that had saved possibly hundreds of thousands of newborns from HIV infection and early death. At stake was the future of one of the few successful programs for intervening in the runaway world AIDS crisis. Or was it?
HIV/AIDS is a slow but persistent disaster, washing over the continents of Africa and Asia with the deadly impact of a new tsunami every few weeks, killing 8,000 people per day, leaving decimated families and legions of orphans in its wake. Children are especially hard hit. An infected infant in a resource-poor setting will likely suffer from intestinal illness and poor nutrition and will generally lack access to pediatric formulations of antiretroviral (ARV) medicines, even in the few places where drugs are available for the parents. Few live to be five. Without treatment, about one in four infected mothers will deliver an infected infant. A baby fortunate enough to be born without HIV runs a high risk of acquiring the virus during the first few months of life if breast-fed by an untreated mother. With nearly 2,000 infants infected each day, the search for practical interventions to reduce this number has been a passionate quest.
One of the most important early government-sponsored HIV studies investigated using an ARV during pregnancy to reduce the risk of a mother passing the virus to her child. The trial, ACTG 076, reported its dramatic findings in 1994: using AZT during the last six months of pregnancy and during the first six weeks of the infant's life could reduce mother-to-child transmission (MTCT) of HIV by nearly two-thirds. The results immediately changed the standard of care for women with HIV in the U.S. and Europe and rates of infections at birth soon plummeted. But the AZT regimen called for diagnosing the mother's HIV status and starting treatment as part of routine prenatal care. What about women who do not have access to pre-natal care? What about women with HIV who only appear for medical attention on the day they go into labor? What about women who give birth at home?
About half of mother-to-child HIV infections occur during labor and delivery; about 15 percent to 20 percent occur earlier in the pregnancy; and the rest occur through breast-feeding. Research has shown that the risk of transmission during labor and delivery is strongly associated with the mother's viral load, and that reducing viral load temporarily -- for example with a single dose of nevirapine -- can reduce the risk by about half. Other drug regimens with longer dosing durations can reduce the risk further and regimens that extend therapy to the mother and baby during the initial months of breast-feeding can lower infection rates even more.
In 1997, a study was designed to test a drug that could be given as one pill to the mother during labor and one pill to the baby, for use in situations where the standard regimen was not feasible. The drug to be studied, nevirapine, was selected because it was quickly absorbed and distributed throughout the body (including the neonate in utero and breast milk); because it could rapidly lower levels of detectable virus circulating in the blood; and because it had a long half-life in the body, meaning it could sustain its effective concentration throughout labor and delivery. Importantly, nevirapine, when tested in pregnant animals, had not been observed to cause birth defects. This strongly distinguished it from efavirenz, a drug in the same class as nevirapine with similar pharmacokinetic properties, but which had produced a high rate of severe birth defects when monkeys were exposed during early pregnancy.
Because AZT had made mother-to-child transmission of HIV rare in the developed world, it would be unethical and impractical to test single-dose nevirapine where proven methods were available. Since the greatest need for a simpler regimen for prevention of mother-to-child transmission (PMTCT) was in Africa, a "proof of concept" study was designed by researchers from Johns Hopkins University to compare single-dose nevirapine with a short course of AZT in 600 mothers in Kampala, Uganda. The trial was called HIVNET 012.
It was a great day when interim results were published in 1999 showing that single-dose nevirapine could reduce MTCT by half during the first 14-16 weeks in a breast-feeding population at a cost of about $4.00. Soon, a crusade was launched to bring the drug to every village and township. But nothing is as wonderful as it seems and the rollout of PMTCT has been painfully slow. It is estimated that less than five percent of women who will need PMTCT treatment in 2005 will have access to it.
Continuously dosed nevirapine, however, used for chronic therapy for HIV infection, is another story. As blood levels of nevirapine build up with continuous dosing, some people, perhaps five percent, will have a reaction to the drug in the form of a rash or liver problems. In a small proportion of those who have reactions, if the drug is not stopped, the result can be deadly. In the words of the FDA:
"Severe, life-threatening, and in some cases fatal hepatotoxicity, including fulminant and cholestatic hepatitis, hepatic necrosis and hepatic failure, has been reported in patients treated with VIRAMUNE [nevirapine]. These events are often associated with rash. Women, and patients with higher CD4 counts, are at increased risk of these hepatic events. Women with CD4 counts >250 cells/mm3, including pregnant women receiving chronic treatment for HIV infection, are at considerably higher risk of these events."
These rare but serious complications of chronic nevirapine therapy came to light after a series of tragic events, including a few health care workers who took nevirapine for post-exposure prophylaxis after a needle stick accident and subsequently died or required liver transplant. Typically, the victims were women. Part of the confusion arising from the recent news reports about the NIH "cover up" was the report of a woman in a NIH-sponsored clinical trial who continued to receive nevirapine after blood tests showed that she was having a reaction. She subsequently died. And the difficult climate for treatment in South Africa was complicated by the nevirapine-associated deaths of two women in a clinical trial there in 2000.
Although single-dose nevirapine does not lead to these kinds of serious reactions, a smoldering problem with drug resistance has started to flare up. Currently available NNRTI drugs, which include nevirapine and efavirenz (trade named Sustiva or Stocrin) have a low genetic barrier to resistance and HIV can quickly find an escape mutation if allowed to replicate in the presence of sub-therapeutic concentrations of these drugs. The problem with single-dose nevirapine is that the very long half-life that makes it ideal for reducing viral load during labor and delivery, also gives HIV a chance to start spinning off drug resistant mutants as the drug's concentration slowly tapers off. Some studies suggest that over half the women who take single-dose nevirapine may develop resistant HIV, thereby compromising their chances to someday benefit from continuous nevirapine therapy for their own health. In the starkest terms, some say that saving the baby means sacrificing the mother.
Among prevention fundamentalists these arguments take on a kind of pro-life versus pro-choice tinge, with one side championing the most innocent victims at all costs and the other protesting the treatment of women as no more than incubation vessels, consigned to the graveyard when their genetic duty is done. Most researchers involved in these issues, though, are looking for practical solutions. They admit that single-dose nevirapine has limitations but see it as the thin end of the wedge that helps a treatment program get on its feet, gives staff experience with antiretroviral drugs, provides an incentive for voluntary counseling and testing and can be replaced with more demanding regimens once capacity is built and trust is established with the community. Already, programs such as MTCT-Plus, which aims to treat the parents as well as the child, are widely considered much more sustainable solutions.
New ideas are emerging for improving the controversial regimen, such as covering the "tail" of dwindling nevirapine concentration with a few days of another antiretroviral drug, such as Combivir. But it's also becoming clear that the investment in infrastructure required to make even simple single-dose nevirapine work as well as it does in a clinical trial setting, is just as well suited for introducing more effective and less problematic regimens. This field is constantly in flux and with new data expected in 2005, the consensus about the minimally acceptable PMTCT regimen may be expected to shift.
Although the conditions under which the trial was conducted were not necessarily unusual for that setting and may have been adequate for the limited purpose the trial set out to achieve, HIVNET 012 was never intended to provide data to support a submission to the fastidious FDA. It became clear that to obtain approval the sponsor would have to conduct another trial to back up the first. This was judged not practical and the application quietly withdrawn.
Despite the sketchy safety database in the Kampala study, audits by the FDA, Boehringer-Ingelheim and by the NIH all agreed that single-dose nevirapine did prevent mother-to-child transmission of HIV. Furthermore, there were no suggestions of safety problems in any other studies of single-dose nevirapine. Given the delicate status of PMTCT programs in South Africa, officials at the NIH felt it was important to keep the message straight: while HIVNET 012 did not supply the kind of pristine data that the FDA requires to approve a drug in the U.S., this did not mean the conclusions were flawed or that the drug was unsafe when used for PMTCT.
One of the problems the FDA found was that reports of adverse events in the study case files could not be confirmed with original hospital records. An investigator familiar with the trial site at Mulago Hospital has described the hospital's records as stacked along hallways and down a staircase in public areas, noting that these conditions were not unusual for hospitals in the developing world. It seems clear there was avoidable sloppiness in the research practices in HIVNET 012, but there were also practical limits to how strictly the investigators could adhere to ideal procedures. If the FDA's standards are to be met, in effect it will mean recreating much of the infrastructure of Western hospitals at every African site. But what would be the point of such research, some wondered, if the conditions of the research were so completely divorced from the realities of the regions they were operating in? How meaningful would the results be? More compelling perhaps is the thought of how many people would die if research stood still while waiting for the FDA's standards to be reached.
The NIH is in the process of improving its capacity to do HIV research in developing world settings by retooling its clinical trials networks to begin cooperating with international sites. On one level, this has been criticized as a "Willie Sutton" ploy, an attempt to sustain the flow of grant money to U.S. investigators by taking the research to where the patients are. Some also worry that building a high-tech research infrastructure in resource-poor regions is not a sustainable solution; that it will be vulnerable to collapse when U.S. interest inevitably shifts to other projects. But the main investment in building a quality clinical research organization is in establishing consistent procedures for performing routine medical tests and recording the data; ultimately it is about training, education and building the capacity of the staff. This may be what it takes to satisfy the FDA's standards, but achieving this level of quality is not cheap or easy. Furthermore, macro-economic factors in the developing world have tended to make retaining skilled staff difficult when there are creditor mandated caps on health care spending by governments and Northern markets for cheap skilled labor that lure employees away as quickly as they are trained.
The Kampala research site had been closed for a year and a half while the audit was underway and was preparing to reopen about the time that Fishbein took over his position as director of the newly created Office for Policy in Clinical Research Operations. One of his first acts on the job was to question whether the site was really ready to reopen and he asked Edmund Tramont, the chief administrator of DAIDS, to let him establish his department's authority by claiming the responsibility to make the decision. In an email made public by Fishbein, Tramont implies Fishbein's role is advisory and states that he wants the Kampala site opened "ASAP because the site is now the best in Africa run by Black Africans."
Fishbein remained critical of the Kampala site and began pressing to reopen an investigation of the flaws in HIVNET 012. Tramont expressed his frustration in a brief email to the DAIDS staff, "Folks, HIVNET 012 has been reviewed, re-monitored, debated and scrutinized. To do anymore would be beyond reason. It is time to put it behind us and move on."
As is evident from internal memos released by Fishbein, Tramont was concerned about a pending visit by President Bush to the Kampala site during an African tour to promote his Emergency Plan for AIDS Relief (PEPFAR). Tramont might also have been concerned about the fragile state of NIH's capacity building efforts in Africa and about the potential damage to emerging PMTCT programs in South Africa and elsewhere if the situation was not carefully managed. Jonathan Kagan, Tramont's deputy, advised him to let Fishbein conduct his investigation lest it appear that it "look like we have something to hide." Furthermore, Kagan wrote, "We should NOT be motivated by political gains. ..."
At some point during the initial 16 months of his employment, Fishbein apparently learned that he was to be fired before his two-year probationary period ended. To protect his job (at a reported salary of $178,000 per year -- stratospheric by NIH standards) Fishbein invoked the Federal Whistleblower Protection Act, intended to shield employees who uncover government waste, fraud and abuse. He claimed he was the victim of retaliation after refusing to go along with a cover-up of the flawed study, which, he later testified, could have "fatal implications." An NIH spokesperson said he was being fired for "poor performance." Fishbein's bid for whistleblower status was denied by an administrative law judge in November and he went public with his charges in a series of Associated Press articles in mid-December. Fishbein says he is simply trying to defend his reputation and that his exposure of the cover-up at NIH should be seen as a public service. He makes this case on his website, HonestDoctor.org, with a list of 19 allegations, including sexual harassment and bid rigging.
When the story broke, reaction was swift. Senator Charles Grassley called for a Justice Department Investigation. Jesse Jackson called it a "crime against humanity" and said, "Research standards and drug quality that are unacceptable in the U.S. and other western countries must never be pushed onto Africa. We should stop discounting the lives of Africans." Most ominously, the South African ruling party, the African National Congress (ANC), published an attack on the NIH accusing them of conspiring with Boehringer-Ingelheim to promote the sale of nevirapine in Africa, treating Africans like "guinea pigs." The article was unsigned, but President Thabo Mbeki was assumed to be supportive of the paper's conclusions. The ongoing PMTCT programs in South Africa have been mandated by the courts, but it remains unclear what impact this episode may have on their future.
Lies and misinformation are difficult to call back once launched, especially when they are tied to inflammatory rhetoric. Unfortunately, the fallout from Fishbein's bid to keep his job may end up killing far more people than nevirapine ever will.
|From a Statement by the Treatment Action Campaign (TAC), South Africa|
It is critical that public confidence in a life-saving treatment used in many public facilities throughout South Africa and the developing world should not be undermined without due cause. Therefore, the TAC points out the following: