[Skip to Content]
[Skip to Content Landing]
Views 11,866
Citations 0
Medical News & Perspectives
March 18, 2020

Confirmatory Trial for Drug to Prevent Preterm Birth Finds No Benefit, So Why Is It Still Prescribed?

JAMA. Published online March 18, 2020. doi:10.1001/jama.2020.1486

Preterm birth represents a major US public health problem, with 1 in 10 neonates born before 37 weeks’ gestation in 2018, placing them at a greater risk of death and disability.

So it’s not surprising that pregnant women who’ve already spontaneously delivered a baby preterm would be eager for a treatment that promises to help them get closer to term. Same goes for their physicians.

For nearly a decade, that treatment has been Makena (hydroxyprogesterone caproate), a long-acting form of a naturally occurring progesterone. Makena was approved in February 2011 under accelerated approval regulations of the US Food and Drug Administration (FDA),which allow promising drugs to enter the market based on a surrogate end point benefit—in this case, reducing the risk of delivery before 37 weeks—that is likely to predict a clinical benefit.

Fast-track approvals are granted on the condition that drugs’ sponsors conduct additional studies to confirm a clinical benefit, such as reduced neonatal morbidity and mortality. But what happens when the confirmatory study fails to confirm the effectiveness of a drug that has become the standard of care?

That question is dogging Makena. In March 2019, AMAG Pharmaceuticals, for whom Makena was its best seller before sales dropped in 2019, announced that the confirmatory trial, called PROLONG (Progestin’s Role in Optimizing Neonatal Gestation), found that the drug was no more effective than a placebo.

Four days after publication of the PROLONG results in late October, the FDA convened an advisory committee meeting to help it decide whether to keep Makena, injected weekly beginning at 16 through 20 weeks’ gestation, on the market. The panel evaluated comments from women who credited Makena for their full-term deliveries and AMAG representatives’ explanations of why PROLONG’s findings differed substantially from the 2003 trial on which Makena’s approval was based.

After weighing the evidence, the panelists discussed and voted on questions posed by the FDA. They voted unanimously that the PROLONG trial did not “verify the clinical benefit of Makena on neonatal outcomes.” Of the 16 panelists, 13 voted no and only 3 voted yes when asked whether they thought PROLONG and the drug’s original trial provided “substantial evidence of effectiveness of Makena in reducing the risk of recurrent preterm birth.”

However, the vote on whether to withdraw Makena was much closer. Nine committee members said it should be withdrawn and 7—including 5 of the 6 practicing obstetricians on the panel—said it should remain on the market, under the condition that AMAG conduct another confirmatory trial. Whether such a trial is feasible isn’t clear, though. Several panelists noted that pregnant women who’d already delivered preterm would be unlikely to enroll in a study in which they might get a placebo instead of the only medication indicated for reducing the risk of another preterm birth. None of the panelists voted for a third option: leaving Makena on the market without requiring a new confirmatory trial.

“[W]e are committed to working with the FDA to identify feasible ways to generate additional efficacy data on Makena while retaining current access to the therapy for at-risk pregnant women,” PROLONG coauthor Julie Krop, MD, AMAG’s chief medical officer until her March 31 departure, said in a prepared statement after the advisory committee vote.

For now, Makena remains on the market, and the Society for Maternal-Fetal Medicine (SMFM) and the American College of Obstetricians and Gynecologists (ACOG)—the 2 main professional organizations of physicians who care for women at risk of delivering preterm—continue to recommend that members prescribe it to patients for whom it is indicated.

At least a few of the groups’ members have questioned whether AMAG’s role as a major donor to SMFM and ACOG (it’s also a March of Dimes “corporate partner”) has affected the organizations’ judgment when it comes to Makena. And at least a few maternal-fetal medicine specialists said PROLONG confirmed what they had already suspected, given that they considered the original trial leading to Makena’s approval to be flawed.

The Meis Trial

Hydroxyprogesterone caproate was introduced in the United States in 1956 under the brand name Delalutin for abnormal bleeding in women with uterine cancer, menstrual disorders, and miscarriage.

A 1990 meta-analysis concluded that although progesterone did not appear to prevent miscarriage, it might protect against preterm birth.

Delalutin, which hadn’t been marketed for several years, was withdrawn in 2000 for reasons other than safety or efficacy, but interest in hydroxyprogesterone caproate revived in 2003. That’s when a double-blind, placebo-controlled trial found that the drug significantly reduced the risk that pregnant women who’d had a spontaneous preterm singleton birth would have another. Compounding pharmacies began making hydroxyprogesterone caproate and selling it for about $15 a shot (which is why Makena’s initial price of nearly $1500 per shot made headlines nationwide).

Dubbed “the Meis trial” for lead author Paul Meis, MD, of Wake Forest University, the hydroxyprogesterone caproate study was conducted by the Maternal Fetal Medicine Units Network of the National Institute of Child Health and Human Development (NICHD), one of whose stated goals is “to provide the rationale for evidence-based, cost-effective obstetric practice.”

At 19 US sites, the trial enrolled 463 pregnant women who’d previously had a spontaneous preterm delivery. The trial had planned to enroll 500 women, but, as Meis, now a professor emeritus, said in a Wake Forest news release, “the evidence of this treatment’s effectiveness was so dramatic, the research was stopped early.”

The drug reduced the risk of delivery at less than 37 weeks, less than 35 weeks, and less than 32 weeks. Infants of treated women had significantly lower rates of necrotizing enterocolitis, intraventricular hemorrhage, and need for supplemental oxygen than those of women in the placebo group.

Oddly, though, the percentage of preterm births in the placebo group was much higher than expected—54.9% delivered before 37 weeks. A 2017 meta-analysis involving 55 197 women found the absolute risk of a recurrent spontaneous preterm birth was 30%—6.3 percentage points lower than the Meis trial’s treated group. The Meis trial authors attributed its high rate of preterm births to participants’ previous preterm deliveries, which on average were at 31 weeks’ gestation and, in a third of women, numbered 2 or more, putting them at high risk of a recurrent preterm delivery. But some skeptics speculated that the trial wasn’t adequately randomized, so hydroxyprogesterone caproate appeared to be highly effective only because the control group fared so much worse than most women who’d had a prior preterm birth.

Lisa Kammerman, PhD, the FDA biostatistician who helped assess the Meis trial when the agency was considering Makena’s approval, found it insufficient to support the drug’s effectiveness. “There were just too many contradictions and inconsistencies for a single study to stand on its own,” Kammerman, who left the FDA in 2014 and now works as a consultant, told JAMA.

For example, she noted that among the trial’s 19 sites, the University of Alabama at Birmingham had enrolled 27% of all the participants. What’s more, the Birmingham center accounted for 44% of women who were enrolled at 18 weeks or earlier, so other sites had disproportionately few women enrolled at the same pregnancy stage, Kammerman wrote in her final evaluation of the Meis study. Those disparities suggested confounding of study sites and gestational age at randomization.

In addition, the trial didn’t identify the optimal time to start the drug, even though it appeared most effective when begun at 18 weeks’ gestation or earlier but ineffective when started at 20 weeks or later, Kammerman wrote. However, she said, the rate of fetal and neonatal deaths was most pronounced among women who started the drug at 18 weeks or earlier. Overall, fetal and neonatal deaths occurred earlier and were more common among women treated with the drug than those who received a placebo. Nearly 4% of the 306 women in the treated group had a second-trimester miscarriage or stillbirth, compared with about 1% in the placebo group.

“Meis was never designed for drug approval,” said Kammerman, citing the questionable randomization and the lack of information about treatment timing. Besides, as Meis noted in the Wake Forest news release, hydroxyprogesterone caproate was already readily available.

But in 2006, Adeza Biomedical used the Meis data to support an application for FDA approval of Gestiva, a brand of hydroxyprogesterone caproate, for the prevention of preterm birth. By the time it was approved, the drug’s name had changed to Makena, and a few years later, AMAG became its manufacturer. (In 2018, when the first generic intramuscular hydroxyprogesterone caproate injections entered the market, AMAG introduced the Makena subcutaneous autoinjector, on which it holds a patent until 2036. However, both Meis and PROLONG studied intramuscular administration, not subcutaneous.)

Why Should It Work? No One Knows

After the Meis trial was published, a group of payers and maternal-fetal medicine specialists, including March of Dimes representatives, met at the US Centers for Disease Control and Prevention to discuss hydroxyprogesterone caproate.

Michal Elovitz, MD, was among them. At first, Elovitz said, enthusiasm for the drug was so great that the committee discussed treating every pregnant woman with a prior spontaneous preterm birth, unless she opted out. “We don’t do that with any other drug,” noted Elovitz, director of the Maternal and Child Health Research Center and the Prematurity Prevention Program at the University of Pennsylvania.

But, she recalled, someone at the meeting asked, “How is giving a little bit of progesterone, when there’s already a sea of progesterone [in the uterus during pregnancy] going to help?”

Elovitz has tried to answer that question in her research, but without success. The problem, she said, is that “we don’t understand the pathophysiology of preterm birth.”

“I don’t have a good argument for why it works,” she said of Makena. “There should be a clear biological mechanism.” Sean Blackwell, MD, PROLONG’s lead author and former SMFM president, said he also believes that “we don’t know why it works, if it works,” but that’s true of “about half of the medications we use during pregnancy.”

And that’s a problem, Elovitz said.

“There aren’t enough studies in women’s health. There aren’t enough studies in pregnant women. We hate doing nothing because there’s not enough good evidence to do something,” she said. “There seems to be this bias of so desperately wanting this drug to work. A bias based in beneficence, not in ignorance. Unfortunately, that rolls over into paternalism.”

Meanwhile, 2 New York women—who had 5 preterm births between them despite taking Makena—in November filed a class action lawsuit against AMAG in US District Court, alleging that the company continued to market the drug even though it knew long before the final PROLONG results became available that it didn’t work. According to the lawsuit, the women paid out-of-pocket for Makena, “which is exorbitantly priced and is painful to take.”

Makena Didn’t PROLONG

Makena’s PROLONG confirmatory trial began in 2009, before the drug even came on the market. According to AMAG, it was the largest trial of any progestogen for preterm birth prevention. In it, 1130 women were randomized to Makena and 578 to placebo at 91 sites, 52 of them outside the US. About 77% of participants were from outside the US, “which was necessary given the routine use of [Makena] as the standard of care” in the US, according to AMAG.

Women were enrolled from the first day of their 16th week of gestation through the 6th day of their 20th week. Preterm births before 35 weeks occurred in 11% of women in the treated group and, similarly, in 11.5% of those in the placebo group. Both rates were far lower than the 20.6% and 30.7%, respectively, reported in the Meis trial.

Makena’s proponents blame demographics, not the drug, for why PROLONG’s results were so different from the Meis trial findings. They argue that PROLONG participants, largely from Eastern Europe, started out with a lower risk of recurrent preterm birth than the US women in the Meis trial.

“The vast majority are from the Ukraine and Russia,” said PROLONG coauthor Tracy Manuck, MD, medical director of the Prematurity Prevention Program at the University of North Carolina, Chapel Hill. “They’re married. They have good social support. The patients are significantly different.”

Indeed, a 2012 World Health Organization report found that, overall, preterm birth rates were far lower in Ukraine and Russia than in the United States.

And yet, while Makena didn’t lower the risk of preterm birth before 35 weeks in Ukrainian women, it did in Russian women, according to PROLONG supplementary material. The treated Russian women’s risk was 2 percentage points lower than that of their countrywomen who received a placebo shot—the same difference seen among US PROLONG participants.

Another reason for the discordant findings in the Meis and PROLONG trials could be their racial makeup, AMAG said. In its briefing document for the FDA advisory committee, the company pointed out that 59% of Meis participants were black and that studies show black women have twice the risk of delivering preterm, or before 37 weeks, compared with white women. But in PROLONG, only 1 of the women from outside the US and 29% of the US women were black. However, as Elovitz pointed out, ”There’s no sub-analysis of Meis that argues it only works in black women,” which seems to shoot down the notion that Makena looked ineffective in PROLONG because the vast majority of participants were white.

“Studies aren’t perfect,” acknowledged Blackwell, who argued at the 2020 SMFM meeting that PROLONG doesn’t answer questions about Makena’s use by US women—although that was the exact reason it was conducted. His conclusion: “Research is messy.”

It Couldn’t Hurt?

Most of the discussion about whether Makena should stay on the market has centered on its effectiveness, not safety.

“If there are individuals that it works for, and we’re able to keep some babies out of the NICU [neonatal intensive care unit],…it doesn’t appear that we’re doing that at the cost of harm,” Manuck said.

Although the Meis trial found an increased risk of stillbirth and neonatal death among women who began treatment before 18 weeks, the PROLONG trial did not.

However, a recent meta-analysis concluded that women who receive hydroxyprogesterone caproate to prevent preterm birth have a higher risk of gestational diabetes. Physicians should consider becoming even more vigilant for gestational diabetes in women treated with the drug and reconsider prescribing it for patients who have gained excessive weight during pregnancy or have polycystic ovary syndrome or a family history of diabetes, the authors advised.

Some observers have questioned whether hydroxyprogesterone caproate might have as yet unidentified long-term adverse effects in women or their children.

The group DES Action USA started a Change.org petition urging the FDA to withdraw Makena. The organization advocates for women who were exposed to diethylstilbestrol (DES), a synthetic estrogen, and for their families and physicians.

Beginning in the 1950s, DES was given to millions of pregnant women in the US and elsewhere under the mistaken belief that it prevented miscarriages. In light of a 1971 report linking maternal use of DES during pregnancy to an increased risk of vaginal adenocarcinoma in daughters, the FDA issued a Drug Bulletin in 1972 warning physicians about possible toxic effects. However, DES remained on the market, and some physicians continued to prescribe it through the 1970s.

“As a society, we got lucky with DES because it caused an abnormality that was so weird it was easy to link it,” said Adam Urato, MD, chief of maternal-fetal medicine at the MetroWest Medical Center in Framingham, Massachusetts. “But if [a drug] is causing problems that are going to be less quirky, you’ll miss it.” Such problems might include behavioral or reproductive issues in offspring of women treated with Makena while pregnant, he explained. “If there’s not evidence of benefit, you don’t use it. You don’t risk exposing a generation of fetuses to a synthetic hormone.”

The FDA classifies Makena as Pregnancy Category B, or having no known adverse effects on the fetus. A follow-up to the Meis trial, submitted as part of the application for FDA approval, found no significant differences in health status; physical examination, including genital anomalies; or neurodevelopment between 194 children exposed to hydroxyprogesterone caproate in utero and 84 children born to women in the placebo group. The children’s average age was 48 months, but it could take much longer for potential abnormalities to emerge. For example, daughters exposed to DES in utero have been diagnosed with a rare vaginal cancer in their late teens and early 20s.

Bad Optics?

Over the years, AMAG Pharmaceuticals has provided financial support to the March of Dimes, ACOG, and SMFM, none of which has yet provided guidance that reflects the PROLONG trial’s findings of no benefit. Whether that’s because they get AMAG funding isn’t known—Blackwell emphasized that no one on the SMFM publications committee, which writes guidelines, personally has received AMAG money—but the optics aren’t great. The situation raises questions about whether medical professional organizations should take money from companies that stand to profit from members’ prescribing habits.

“The medical societies in this area are compromised because they accept funding from the pharmaceutical company that makes Makena,” said Michael Carome, MD, director of the Public Citizen Health Research Group, which has petitioned the FDA to withdraw Makena. Urato cosigned the petition.

In 2018, AMAG became a March of Dimes “strategic corporate partner,” funding 9 new Supportive Pregnancy Care sites. Spokeswoman Christine Sanchez said AMAG funding hasn’t influenced her organization’s efforts to prevent preterm births. “We believe patients should always speak with their doctors about their specific health needs,” Sanchez said in an email.

At SMFM’s 2019 annual meeting in Las Vegas, Urato said he was “astounded by how much AMAG Pharmaceutical advertising was everywhere.” He noted that pink mats promoting Makena were scattered on the floor beneath research posters exhibited at the conference.

In the meeting program, AMAG was listed along with only SMFM as a top 2018 donor—having contributed $100 000 or more—to SMFM’s foundation, which provides education and research grants. The program lists AMAG as one of 3 “premier” members of the SMFM foundation’s corporate council. In addition, AMAG funded the development of SMFM’s preterm birth app, and AMAG and SMFM jointly award health policy grants for up to $25 000 each.

The program for the 2020 SMFM meeting in February, at which Blackwell spoke on a panel about hydroxyprogesterone caproate entitled “Where Do We Go From Here?” lists AMAG as the sponsor of a “diversity reception,” one of about a dozen corporate sponsors of a fundraiser for the SMFM foundation, and a premier member of the foundation’s corporate council.

AMAG also was a gold-level sponsor of ACOG’s 2019 annual meeting. ACOG spokeswoman Megan Christin said in an email that the organization “maintains relationships with a variety of organizations, foundations, and corporations committed to fostering excellence in women’s health care, including AMAG. The college accepts financial and in-kind support only if such acceptance does not pose a conflict of interest and in no way impacts the College’s objectivity, priorities and actions.” Corporate funding represents “a small fraction” of ACOG’s operating budget, Christin added.

“I’m a transparency person,” said Blackwell, assistant dean for health care quality in perinatal medicine and women’s health at the University of Texas Medical School at Houston. “If we’re willing to partner with industry, we should be on the up-and-up.”

“A Pretty Big Practice Shift”

Although PROLONG hasn’t spurred SMFM or ACOG to advise against prescribing Makena, discussions at the University of Pennsylvania led to a consensus that PROLONG was a negative trial, Elovitz said. “We wrote guidelines that this should not be routinely offered anymore,” she explained.

Mark Alanis, MD, a maternal-fetal medicine specialist in Colorado Springs, Colorado, said he was puzzled by the latest practice advisories on Makena. “This is the first time that I can remember that SMFM and ACOG have disregarded the conclusions of a large, well-designed multicenter randomized controlled trial in favor of a smaller, more controversial, and less well-designed study,” he said.

Even so, Blackwell said, “There are scientists who believe that Makena works and that PROLONG was an outlier. If we did the PROLONG study first and then we had Meis, we wouldn’t be talking about withdrawing this.”

On the other hand, “if the PROLONG trial data had been available back in 2010…there’s no way they [the FDA] would have approved the drug,” Carome said.

Blackwell said it could be argued that ACOG and SMFM shouldn’t yet advise physicians to stop prescribing Makena, since it’s still on the market. “They can’t say, ‘Stop it,’ and then change their mind again. It’s a pretty big practice shift, and they want to be very thoughtful.” However, he said, it would be another matter if PROLONG had found evidence of harm or physicians had a clear alternative to Makena.

Blackwell said he hasn’t yet stopped prescribing Makena, although he and his colleagues now discuss the PROLONG results when counseling patients about the drug. “I certainly have less certainty than I did before PROLONG,” he added.

×