Skip directly to content

Down in Flames

tagline 1998

Crypto Treatment Hopes Dashed as Sloppy, Divisive Unimed Fails to Provide Clean Data -- or Proof of Efficacy

Unimed's ugly ultimatum

On May 6th, the FDA’s Antiviral Drugs Advisory Committee met to consider an application for approval of Unimed Pharmaceuticals’ cryptosporidiosis treatment, nitazoxanide (NTZ), for the treatment of cryptosporidial diarrhea in PWAs. TAG’s Laura Morrison was there and prepared this report.

Despite more than ten years of clinical trials on dozens of agents for cryptosporidiosis, there is still no FDA-approved therapy and no standard of care for its treatment. In fact, prior to NTZ, not a single New Drug Application (NDA) had been filed for the treatment of cryptosporidiosis. Efforts to find useful therapies have been impeded by the absence of an in vitro model and the lack of a universally accepted animal model with which to screen potential agents. Potential treatments at present (see table below) include spiramycin, diclazuril, letrazuril, octreotide (Sandostatin), azithromycin, hyperimmune bovine colostrum, hyperimmune egg yolks (IGX-CP) and paromomycin (humatin). But nothing is reliably effective, and even the best response rates are notoriously low. Rifabutin and clarithromycin may have a prophylactic effect against C. parvum, but more research is needed. Because of this longstanding void and the devastating, debilitating nature of cryptosporidiosis, Unimed's NTZ application generated an unusual amount of excitement among treatment activists.

Unimed's application itself was unusual in that it was based solely on open-label, compassionate-use data. While Unimed did support a study of NTZ vs. placebo (ACTG 336), that study had to be terminated due to abysmal accrual (only 9 patients over a year's time). The FDA, recognizing the difficulties of enrolling such a study in the era of protease inhibitor combination therapy, encouraged Unimed to submit its NDA based on data collected from its open-label studies. The FDA proposed using data from the placebo arms of ACTG 192 (paromomycin) and Pfizer study 143 (azithromycin) as historic controls.

Unfortunately, even within the FDA's rather loose requirements, Unimed came up far short. Unimed presented data on 226 people with AIDS in three open-label studies. UMD-95-004 was a phase I/Il dose escalation study (500-2,000 mg/day) in 28 PWAs with microbiologically confirmed cryptosporidial diarrhea. Approximately 32% of the patients achieved a complete response (clinical resolution of diarrhea to 1-3 bowel movements per day), and 18% had a partial response (at least a 50% reduction in the average number of daily bowel movements but still >4 bowel movements per day or change in stool consistency to >75% formed at the end of the study). UMD-95-009A and UMD-95-009B were open-label, multicenter studies of 139 and 57 PWAs, respectively, with diarrhea "attributed to" cryptosporidiosis who were refractory to other experimental cryptosporidiosis therapies. Although Unimed documented that approximately 50% of patients achieved a clinical response in terms of reduced watery and total stools, both studies were fraught with problems that limited the credibility of their findings, including:

  • Only 39 patients from 009A and none from 009B met Unimed's definition of "per protocol," i.e., they were on a 1,000 mg NTZ/day regimen, had microbiologically confirmed presence of C. parvum in their stool at baseline, had >4 liquid stools/day at baseline and at least 1 post-baseline visit with stool frequency data.
  • For most patients there was no information about relevant concomitant medications (i.e., antidiarrheals), and there was no analysis of outcome based on whether patients were on a protease inhibitor.
  • Unimed performed its analysis using a "Last Non-missing Observation Carried Forward" (LOCF) method, thus true, full-course outcomes were not known for many patients.

TAG was concerned with these problems, yet we saw that NTZ had an excellent safety profile (minimal toxicity) and apparent activity in at least some patients. Taking all that into consideration (and despite the marginal quality of data Unimed presented), TAG felt NTZ showed an undeniable clinical benefit with few adverse effects. Thus we drafted a consensus paper, signed by more than 20 community groups nationwide, that concluded, "in a risk-benefit ratio, NTZ's modest clinical benefit outweighs the safety risks." We recommended that the FDA approve NTZ for cryptosporidial diarrhea and work with Unimed in developing post-marketing studies to determine the optimal dose and duration of treatment.

Regrettably, it became apparent during the FDA's presentation at the hearing that Unimed's data were even weaker and sloppier than we suspected. The FDA reviewer revealed that 34% of patients in the three trials had been lost to follow up. When asked what had happened to those patients, none of Unimed's representatives could answer. (We later found out that the vast majority of these lost-to-follow-ups had died.)

The FDA reviewer said that if we assumed the worst case -- that those lost to follow up were failures -- clinical benefit would drop to just 31%. Even in a best-case scenario, the FDA calculated a clinical benefit in only 43% of patients.

When an FDA statistician compared 91 patients from the NTZ studies to the 14 patients in the placebo arm of ACTG 192 (and separately to the 41 patients in the placebo arm of Pfizer study 143), no statistically significant difference was observed using numerous criteria. (It should be noted that the small numbers of patients in each placebo arm made achieving statistical significance a particularly high hurdle. Nonetheless, this torpedoed Unimed's chances.)

Perhaps the most damning part of the hearing came when the Advisory Committee members asked their questions. Unimed was grossly unprepared and was not even able to handle a challenge from panel member Cynthia Sears (a gastroenterologist from Johns Hopkins) on the percentage of cryptosporidiosis patients whose symptoms spontaneously remit. The panel members, sensing a weak applicant, went for the kill.

By the time the panel broke for lunch there was little doubt what the outcome would be. They voted 9-1 against approval, with TAG's Michael Marco, the community representative on the advisory committee, casting the lone "yes" vote in deference to the community consensus. The clear message after the FDA hearing was that Unimed would need to do another study, one that could be small but tightly monitored and rigorously assessed. The biggest fear of treatment activists was that Unimed would abandon NTZ for cryptosporidiosis entirely. In fact, the position Unimed struck following its defeat at the FDA was even worse.

In a May 14th conference call Unimed Vice President Dr. Robert Dudley told treatment activists from around the country that Unimed was seeking a "conditional approval" from the FDA that would allow the company to market the NTZ pending completion of a confirmatory study. Barring that, the company would discontinue its cryptosporidiosis research and end its expanded access program as well. Furthermore, Unimed would not grant rights to any other company willing to pursue the NTZ indication. Dudley pleaded poverty saying Unimed had already invested too much in NTZ at the expense of other agents. Then he said the company didn't have the supply of NTZ available to continue expanded access -- this despite the fact that the company is planning to begin studies of NTZ for treatment of H. pylori. Dudley's ultimatum smacked of blackmail. Either the community mounted a gargantuan effort that would somehow sway the FDA to grant conditional [sic] approval (based on nothing) or we would lose access to a promising therapy (short of getting it through buyers clubs which import the drug from Mexico for those with the financial resources to purchase it). TAG sent word to the community that we flatly refused to play into Unimed's hands (or speak to their Ken doll, ex-AIDS activist PR agent) by senselessly pleading with the FDA to conditionally [sic] approve a drug with no efficacy data. Unimed didn't even realize that there is no such thing as "conditional approval" -- it's called "accelerated approval" and it comes with strict FDA regulations.

After considering Unimed's position, many in the community finally decided they were unwilling to make excuses for the company's sloppy data and plead with the FDA. Linda Grinberg and Martin Delaney (of Project Inform) drafted a community consensus statement calling on Unimed to immediately amend its application to seek instead a cost-recovery Treatment IND (investigational new drug) and urging the FDA to immediately begin work with Unimed to design an appropriate follow-up study capable of confirming the drug's efficacy. By June 1, signators to the statement included TAG, Project Inform, FAIR and the San Francisco AIDS Foundation. While the FDA has indicated willingness to offer a cost-recovery Treatment IND, Unimed is apparently unwilling to pursue that option at this time. As TAGline goes to press, treatment activists continue to develop a strategy which might convince an incompetent, mean-spirited little company to do the right thing.

The Elusive C. parvum

Potential TreatmentResponse Rates
azithromycinLittle effect§
bovine colostrumNo effect
*ACTG 116;
Loeb 1995 (69% of responses were 'short-lived');
Romeu 1991;
§lactose free formulation available from Pfizer: 1-800-742-3029;
Fries 1994 (GalaGen's new formula being tested 1-800-372-2437);
**ACTG 192