Science Daily reports on a study persuasively showing that progesterone injections don’t reduce recurrent preterm birth rates. What’s more, they increase the likelihood of developing gestational diabetes.
The study compared rates of preterm birth after 20 weeks and before 35 weeks in 430 consecutive women with prior preterm birth(s) who were given progesterone injections with the rate in women before progesterone treatment was introduced. Investigators matched each treated woman with three women in the historical cohort who shared the same preterm birth history (number and order of preterm vs. term births), ethnicity, and BMI. They found that recurrent preterm birth rates (25% vs. 23%) were similar as was pregnancy duration and plasma progesterone concentrations in women who had and didn’t have preterm births.
In addition, 13% of treated women were diagnosed with gestational diabetes compared with 8% among their matched controls, a side effect also found in another study that looked at this. It is understandable that progesterone would increase likelihood of a gestational diabetes diagnosis because, as study authors write, progesterone promotes maternal resistance to insulin, which increases circulating glucose. (FYI: The function of increasing maternal blood glucose levels is to provide fuel to support the growing baby, and, in fact, mild glucose intolerance in an otherwise healthy woman is no big deal and a situation where the cure is likely to be worse than the disease.)
Progesterone treatment is now widely used in the U.S. in the belief that it has been established as both effective and harmless and fully deserving of its FDA approval. (See the Cochrane systematic review on the subject.) However, this is by no means the case. Meis (2003), the trial on which the FDA based its preliminary approval, was extremely flawed (Keirse 2004). Chief among its flaws was that the improvement seen wasn’t because progesterone treatment reduced preterm births. Preterm birth rates in the treatment group were identical with that anticipated in women not having treatment based on an earlier study of similar women conducted at hospitals in the same medical center network. The difference lay in women in the placebo group having much higher preterm birth rates than anticipated.
Safety remains unproven as well. A follow-up to the Meis (2003) trial evaluated 194 treated and 84 untreated children at around age 4 and concluded that treatment had no ill effects. But this is far too few children to detect differences in uncommon adverse effects. Moreover, the DES debacle should have taught us that age 4 is far too young to evaluate the potential effects of weeks’ worth of exposure to excess amounts of a sex hormone.
These deficiencies notwithstanding, the Cochrane review includes the Meis trial (as well as another trial questioned by Keirse), and, because of its size, its inclusion heavily weights the pooled analysis. As for safety, the Cochrane review merely notes in passing that only two studies report longer-term results in children (the Meis trial at 4 yrs and another trial’s follow-up at 2 yrs) and calls for further research. (Both a strength and a weakness of systematic reviews is that they strive to minimize subjectivity. If a study fits the review’s parameters, it will be included regardless of problems not captured by the inclusion/exclusion criteria, including possible data manipulation. Systematic reviews, including this one, also treat large, well-designed and conducted trials as equivalent to small trials conducted in hospitals in countries such as Egypt, India, or Iran, which at worst, may lack research expertise and at best, may produce results not applicable to high resource countries.)
On top of all this, the entire edifice of progesterone treatment was built without ever ascertaining whether women threatening preterm birth are deficient in progesterone. This violates a research principle that proposed treatment should be biologically plausible, and the current study’s failure to find a correlation between plasma progesterone levels and pregnancy duration suggests that it’s not.
There’s more to this tale that should raise eyebrows. The current study’s authors explain that the FDA granted preliminary approval for progesterone treatment on the basis of the Meis trial and its follow up study but required a larger, confirmatory trial using prevention of preterm delivery at less than 35 weeks’ gestation as a more clinically meaningful endpoint. Once the drug company that had bought the manufacturing rights launched that trial, the FDA granted it temporary approval to market the drug under the brand name Makena. The drug company then promptly announced that Makena would cost $1500 per injection, as opposed to the $15 per dose it cost when produced by a compounding pharmacy, and has doubtless been making a bucket load of money since. The current study’s authors calculate that with 133,000 women with prior preterm births delivering each year in the U.S., if each woman receives 20 injections, this would amount to almost $2 billion dollars. And here’s the kicker: the trial, which began in 2009 and was supposed to end in 2013, has been extended to late 2018, 5 years past its original close date. It kind of makes you wonder why the drug company keeps pushing back the date. Dirty work at the crossroads? You decide.
The worst of it is that we’ve known for a decade about a preventive treatment that holds great promise, is much cheaper than Makena, and has no adverse effects: group prenatal care by midwives (A.K.A. “Centering Pregnancy”). A randomized controlled trial was published in 2007 in Obstetrics and Gynecology—in other words, not exactly buried in an obscure journal (Ickovics 2007). Participants were 1000 women, most of whom were low-income black women, a group at high socioeconomic risk for preterm birth, who were assigned to either group prenatal care or standard care. The preterm birth rate was 14% in the standard care group versus 10% in the group prenatal care group, a one-third risk reduction after controlling for factors that increase risk of adverse perinatal outcomes. In African-American women, the reduction was even more striking: 16% versus 10%. (Group prenatal care, BTW, has biological plausibility. Forming a community, learning self-care strategies, trading advice, etc. reduces stress, and preterm birth is associated with chronic high levels of stress.)
Regardless, this study and others finding that progesterone treatment doesn’t work are unlikely to have any effect (Grobman 2012; O’Brien 2007). For one thing, truthiness, that gut feeling that something must be true because it fits with one’s beliefs, acts as blinders. For another, as we saw above, “There’s gold in them thar injections.” With time, as has happened before, the treatment or test or whatever becomes standard management, no one remembers that there were ever any doubts or that those doubts were never satisfactorily resolved—gestational diabetes, anyone?—and the juggernaut rolls on.