Special Report: 'Stopping a Steam Train'

Immunizations and enrollment in a second trial with MRKAd5 have now been permanently stopped, and volunteers in the Phambili and STEP trials are being unblinded

By Andreas von Bubnoff

Immunizations and enrollment were stopped permanently on October 23 in a second National Institute of Allergy and Infectious Diseases (NIAID)-sponsored trial called Phambili, or HVTN 503, following the recommendation by the trial's data safety monitoring board (DSMB). Both had already been suspended a month earlier in response to the futility analysis by the DSMB of the STEP trial. The Phambili trial was a companion study to the STEP trial testing the same clade B MRKAd5 vaccine candidate at sites in South Africa (see A STEP back?). Phambili's DSMB also recommended that the study investigators unblind all participants, telling them whether they received vaccine or placebo, and counsel them about the possibility of an increased susceptibility to HIV infection due to the vaccine. A similar decision was reached on November 13 for the STEP trial, and now all volunteers in this trial will also be unblinded.

When Phambili was suspended in September, it was at a much earlier stage than the STEP trial. At the time, Phambili had enrolled only 801 volunteers of a planned total of 3,000, 58 of whom had received all three vaccinations. Still, as news of the suspension reached the Phambili trial sites in September, it felt like "stopping a steam train," says Glenda Gray of the Perinatal HIV Research Unit at the University of Witwatersrand and principal investigator of the Phambili trial. At that time, the trial sites were enrolling as many as 50 volunteers a day.

The process of unblinding the Phambili trial began immediately after the DSMB's decision was released on October 23. Gray says it took only 16 days to unblind all participants after the initial announcement. In what she compared to a "military operation," the volunteers were contacted by radio, cell phone short message service (SMS), and phone. All volunteers are still being encouraged to return for study visits. Gray says the decision to unblind volunteers in the Phambili trial made sense because the trial was at such an early stage it would not have yielded any substantial information, even if the participants who were already enrolled were kept blinded.

One goal of the Phambili trial was to see if the candidate vaccine would be effective in areas with HIV subtype C, the most common clade circulating in South Africa. The Phambili trial was also conducted for the most part in heterosexual volunteers and was to enroll 60% women, far more than in the STEP trial, which originally aimed for one quarter of the volunteers to be women. Women are at a very high risk of contracting HIV in South Africa, and while the STEP trial primarily involved men who have sex with men (MSM), the Phambili trial could have helped researchers determine the efficacy of this candidate in women.

When Phambili was suspended it had enrolled approximately 45% women, compared with 38% in STEP. But the HIV incidence rate among women in the STEP trial was very low—only one of the 83 HIV infections through October 17 occurred in a female volunteer. The reasons for this discrepancy are currently unclear. "We don't have enough data from this study to say anything about vaccine effects in women," says Susan Buchbinder of the University of California in San Francisco and principal investigator of the STEP trial.

Following the unblinding of the Phambili trial, there was also some discussion about how the Phambili and STEP trial results would affect other AIDS vaccine trials in South Africa. On November 14, South Africa's Health Minister, Manto Tshabalala-Msimang, said that all HIV vaccine trials would be put on hold, pending review of the STEP and Phambili data, according to Elise Levendal of SAAVI, who was present at a meeting with the Minister that day.

But there are currently no trials in South Africa that have ongoing immunizations, Gray says. A Phase I trial is planned for next year to test DNA- and MVA-based candidates, developed by Carolyn Williamson of the University of Cape Town, in a prime-boost regimen. That trial, Williamson says, is currently under regulatory review with the FDA. "I am hoping that the issues will be resolved so as not to slow down the testing of this vaccine," Williamson says.

The process of unblinding the STEP trial is now also underway, following an announcement from Merck and the HVTN on November 13. An oversight committee comprised of leadership from Merck, NIAID and the HVTN made the final decision, according to Buchbinder, who says this decision was reached after discussing two options: unblinding all volunteers or offering unblinding to all participants while creating a voluntary blinded follow-up arm. Peter Gilbert of the Statistical Center for HIV/AIDS Research and Prevention (SCHARP) in Seattle calculated that if at least 40% of all the STEP trial volunteers remain blinded, the statistical power of the trial would still be intact. But according to Buchbinder there was substantial uncertainty that investigators could learn significantly more from the subgroup likely to remain in blinded follow up. "There were many benefits to unblinding all study volunteers," she says, "including the clarity with which we could deliver risk-reduction counseling messages and for building trust with the study volunteers and the broader community."

Buchbinder says that before the official decision to unblind, some STEP volunteers had already requested unblinding, an option available to all volunteers at any time.

Kristen Jill Kresge contributed reporting to this article.