On May 6, 1998, the Antiviral Drugs Advisory Committee of the US Food and Drug Administration (FDA) reviewed the application by Unimed Pharmaceuticals, Inc., for accelerated approval of nitazoxanide (NTZ) as an orphan drug and voted 9-1 against the approval of the drug for cryptosporidial diarrhea. The promise offered by NTZ had already characterized the agent as "the most promising drug" for the management of cryptosporidiosis in at least one of the current treatment guidelines.1

"It has been very challenging to put this application together," said Mark Goldberger, MD, MPM, director, Division of Special Pathogens and Immunologic Drug Products, Office of Drug Evaluation IV at the FDA. "There is uncertainty in this setting regarding end points and methods of analysis, some of which you will see reflected in the presentation." He asked the committee for guidance on this and future applications dealing with diarrhea in people with AIDS.

Unimed's Robert T. Dudley, PhD, senior vice president, Research and Development, explained that the company had planned to conduct a placebo-controlled trial of NTZ "in treating cryptosporidial diarrhea in individuals with advanced HIV disease [CD4+ <200/mm3]." AIDS Clinical Trials Group (ACTG) Trial 336 was designed to enroll 60 patients. It was closed on May 15, 1998, after enrolling only 10 patients over the course of 15 months.

The problems with enrollment were multiple. The study was designed before the widespread introduction of protease inhibitors and the impact of combination antiviral therapy upon the reduction of the incidence of all opportunistic infections, including cryptosporidiosis. Dudley raised the ethical issue of treating seriously ill patients with placebo, even for short periods, as well as the practical effect. "This ACTG study is a perfect example of patients not being willing to spend three weeks on placebo." Finally, NTZ "has become, on the street, the de facto drug of choice," asserted Dudley. "Its availability is quite widespread, largely because of imported NTZ [by individuals and through buyers' clubs] from Mexico," where it has been approved for use in controlling diarrhea, first for other parasites and later for cryptosporidium.

Dudley described NTZ as "an anti-infective with a broad spectrum of activity against a host of bacteria and parasites. Its mechanism of action is currently unknown. Pharmacologically, NTZ is very rapidly deacetylated by esterases in the plasma, so fast that one cannot actually track the parent compound."

Later, during a discussion with the committee, Dudley acknowledged that it is unclear whether NTZ is absorbed by the parasite while in the gut, "or whether the drug needs to come in from the back side [through blood], or whether the response is due to the fact that over time you're interrupting some life cycle stage that we don't understand."

Rosemary Soave, MD, associate professor of medicine at the Cornell Medical College, was a principal investigator on the study for Unimed. Her decade of research on cryptosporidiosis includes participation in 11 trials involving over 350 patients. She said that the organism was ubiquitous, "but we really don't know the true prevalence of this disease in any specific population."

"Patients who are infected with this parasite have intestines that are literally covered with these organisms," Soave said. "Cryptosporidium maintains a unique relationship with the intestinal cells. It is interiorized by the intestinal cell. The intestinal cell wraps its membrane around it, and the organism maintains an intracellular but extracytoplasmic position. We don't know whether this peculiar location actually protects the parasite in terms of making it inaccessible to any of the chemotherapeutic agents that we administer."

As with some other parasites, it is unclear how cryptosporidium parvum causes disease, but, according to Soave, the effect is a "tremendous outpouring of water electrolytes -- reminiscent of mechanisms such as those caused by enteropathogenic Escherichia coli and cholera toxin." Often the scope and frequency of watery bowel movements leads to dehydration and the need for intravenous replenishment. It may reduce desire to eat and hence contribute to wasting, interfere with sleep patterns, and limit physical activity because of incontinence. All of which contribute to a debilitated state of mind and health.

"It is clear to us that most of the patients with CD4+ counts under 200/mm3 have a version of the disease that is either chronic intermittent, chronic persistent, or progressive/fulminant," Soave said. Recovery may be spontaneous in healthy individuals but seldom in those with low CD4+ counts.

Soave described the difficulties, both ethical and practical, in enrolling severely ill patients (those with "CD4+ counts way below 50/mm3" and stressed by opportunistic infections) "who are on a fast downward curve going towards death" and asking them to first establish a baseline, then perhaps take a placebo for three weeks.

The baseline is crucial, Soave said, because "diarrhea is multifactorial in patients with AIDS. It can be due to concomitant pathogens and concomitant medications." Frequency of bowel movements is easy to record but volume is more difficult to capture.

Soave noted that "in the immunocompetent host, parasitologic clearance occurs at least two to four weeks after clinical clearance. This lag period can be prolonged in patients with AIDS. And, therefore, if you don't extend your study enough, you may not pick up the parasitologic effect."

Soave suspects, "that the lack of therapeutic efficacy in these patients could very well reflect their profound immune defect, as opposed to merely reflecting the absence of drug activity."

The protocol

Dudley reviewed data from Study 009A, a four-week open-label trial containing 139 patients, and Study 009B, a subsequent trial of 56 patients in arms of 1000-mg and 2000-mg doses of NTZ. The studies produced similar results, with no significant differences in response between the two arms. Unimed laid these data over a "historic control" of earlier unpublished data from a placebo arm of a trial of azithromycin conducted by Pfizer (41 patients) and ACTG Trial 192 of paromomycin (18 patients).

Dudley concluded, "Beneficial drug effects for NTZ at 1000 mg were demonstrated in this study by a partial or complete response in approximately 60 percent of patients, significant decreases in liquid and total stools by three to four per day, significant reduction in debilitating effects, improvements in oocyst shedding (48 percent), and stabilization of body weight; and the drug was effective on rechallenge."

Shelley Gordon, MD, PhD, a San Francisco physician, described on behalf of the sponsor an anecdote of a patient whose stools turned negative for c. parvum oocysts while he took NTZ. However, during the regimen the patient also changed antiviral therapy, and while lacking "a clear response in his T cells or viral load," seemed to have a positive response as exhibited by weight gain and a greatly improved sense of well-being. The patient continues to take NTZ as a prophylaxis.

Nancy Stillman, PhD, conducted the FDA's statistical review of the application. She compared "six clinical end points where controlled comparisons were possible" between the NTZ trials and the historic placebos. She concluded, "There were no significant treatment differences observed on any of these six end points."

Committee member Steve Self, PhD, from the Fred Hutchinson Cancer Center at the University of Washington, questioned the last observation carried forward methodology used by Unimed in their analysis. "Missing data typically happens among those whose response is not particularly favorable," he said of the one-third dropout rate. Additionally, such methodology "might actually aggravate the biases that come from missing data-you create trends when, in fact, there are no trends going on."

Cynthia L. Sears, MD, a committee consultant from Johns Hopkins University, questioned the rarity of spontaneous remission with incidents of cryptosporidiosis. She mentioned several papers that documented such events in patients with low CD4+ counts and concluded, "I think it is a clinical pattern that is seen, and we don't understand the reasons for it. It is not common, but I don't know if it is rare."

John D. Hamilton, MD, professor of infectious diseases at Duke University, pushed the sponsor on the precision of diagnosis at baseline. Dudley conceded that in Study 009A only "39 of 139 had incontrovertible evidence" of cryptosporidial infection. Open-label patients were allowed to enroll "if there had been a diagnosis within two months prior to that and the symptoms were consistent with cryptosporidial diarrhea."

James J. Lipsky, MD, a clinical pharmacologist at the Mayo Clinic in Rochester, Minnesota, asked if there had been "any relationship to any pharmacokinetic parameter and efficacy? Any patient outcome for this drug?"

"No," replied Dudley, adding, "Blood samples were not collected from [open-label study] patients at all."

"That is pretty basic," said Lipsky. "There are a lot of unanswered questions there."

After reviewing Unimed's slide of Study 009A showing the correlation of parasitologic and clinical responses, committee chair Scott M. Hammer, MD, Beth Israel-Deaconess Medical Center and Harvard Medical School in Boston, interpreted a limited subset as having "five patients who had a complete clinical response, and who had either no change or a worsened parasitologic response." Hammer saw this as "some evidence of discordance."

During the afternoon discussion, the tenor of the panel became evident: there was little question of safety, but there were many questions surrounding efficacy. "I'm not sure that we would be doing a favor to the community at large to state that efficacy has been demonstrated," said William Christopher Mathews, MD, MSPH, Department of Medicine at the University of California, San Diego.

Hamilton echoed that sentiment. "I would in all conscience be ill-advised to recommend this drug either for a large number of patients with this disease or to an individual."

A coalition of 17 activist community groups had submitted a letter, which acknowledged some of the shortcomings of the NTZ application, but urged its approval nonetheless. Their representative, Michael Marco from the Treatment Action Group in New York, was the only voice on the committee for approval, but it was muted: "A soft 'yes' for approval." The advisory committee voted 9 to 1 against recommending approval to the FDA.

Attempting salvage

In subsequent discussion the committee members indicated the types of information they would like to see in future applications. Henry Masur, MD, from the Critical Care Medicine and Clinical Center at the National Institutes of Health, suggested that "some comparative studies could be done, either looking at different doses" or in tandem with other therapies. His experience has been that "some of the patients who were the most convinced they benefited from other agents were on placebo. So it is very important to have comparative data and regular evaluations and to be able to look over a long time -- six months or one year."

Lipsky stressed the need for a "better understanding of the pharmacodynamics. For instance, does it need to be absorbed? Would you want to promote absorption? How would you rationally dose this drug? That would be crucial to the situation." He did not see the need for placebo-controlled trials; the questions could be studied in a dose escalation trial.

"The Achilles' heel of this study has been a series of methodological issues that either couldn't be addressed or weren't addressed," said Hamilton, "which further suggests that there has been insufficient engagement of several vested interest groups." He singled out the scientific community and those people living with AIDS.

Sears pointed to the study of children in Northeast Brazil who carry cryptosporidium. Asymptomatic carriers have no evidence of inflammation in their stools, while symptomatic children show signs of inflammation. She suggested that intensively studying a small group of patients "to try to develop the correlates to understand who does and does not respond might help us direct future study" of cryptosporidium and NTZ.

Mathews offered that "the study of a relatively small number of patients who are extremely well-characterized would perhaps tease out clearly whether there is any drug effect," possibly within a limited patient subset.

On May 14, 1998, Unimed Pharmaceuticals announced that "the lack of a suitable patient population in which to conduct the type of studies requested by the FDA's Antiviral Drugs Advisory Committee last week means that it will no longer pursue an indication for cryptosporidial diarrhea in AIDS patients." However, in subsequent discussions with the International Association of Physicians in AIDS Care (IAPAC), Unimed stated that they were anxious to explore other options suggested by IAPAC for the resubmission of the NTZ orphan drug request with the data necessary to more adequately evaluate the drug's effectiveness as a treatment for cryptosporidial diarrhea as well as for other parasitic gastrointestinal diseases. A meeting has been scheduled between IAPAC and Unimed at the end of July.

According to Gordon Nary, IAPAC's executive director, "The increasing number of patients who fail antiretroviral therapy will be at greater risk for opportunistic infections. It is therefore essential that more effective treatments be developed for cryptosporidiosis and microsporidiosis. The unique circumstances that marked the submission of the NTZ application to the FDA may have made it impossible to prove whether the agent's promise of efficacy is justified. Unimed's willingness to explore a more structured approach for NTZ's orphan drug status is an important step in meeting the special challenges of cryptosporidiosis and microsporidiosis."

Bob Roehr is a medical writer based in Washington, DC (e-mail: bobroehr@aol.com).


1. Bartlett JG. 1998 Medical Management of HIV Infection, Johns Hopkins University, Department of Infectious Diseases, Baltimore, MD; 1998:p89.