The report by Laughon et al from the National Institute of Child Health and Human Development Consecutive Pregnancies Study is a timely contribution to the ongoing national conversation about how to identify women who might be candidates for progestogen prophylaxis to reduce the risk of spontaneous preterm birth. Algorithms for the use of progestogens have targeted 2 populations of women for treatment: (1) women with a previous spontaneous preterm birth, which is defined as a clinical presentation with preterm labor or preterm premature ruptured membranes and (2) women with a short cervix on transvaginal ultrasound scanning, which is defined as ≤20 mm at <24 weeks of gestation. The literature cited to support these guidelines is based on assumptions that spontaneous preterm birth is distinct from indicated preterm birth and that both are distinct from a short cervix, neither of which may be true. The proposed treatment algorithms recommend weekly injectable 17 alpha-hydroxyprogesterone caproate (17-OHPC) for women with a previous spontaneous preterm birth and daily vaginal progesterone for women with a short cervix, and list no effective treatment for women with a previous indicated preterm birth.

Laughon et al found that the risk of subsequent spontaneous preterm birth is increased in women with a previous indicated preterm birth: “An indicated delivery in a previous pregnancy still had a significant 2.7-fold increased risk for spontaneous preterm birth in the next pregnancy.” The apparent increased risk of spontaneous preterm birth in women with a previous indicated early delivery is likely related to variations in the application of definitions of spontaneous and indicated to clinical data. Both are labels affixed post hoc to clinical presentations rather than to biologically defined pathways. Some labels in each category are obviously distinct (eg, severe preeclampsia [indicated] and progressive cervical effacement and dilation that results in delivery [spontaneous]). But what about a patient who presents with light bleeding? Vaginal bleeding could represent either placental abruption or progressive cervical change. After the fact, the woman with bleeding is likely to be labeled as having had an abruption (an indicated preterm birth) when the true source of bleeding might have been a cervical blood vessel. Women who arrive at the hospital with ruptured membranes uniformly are assigned to the spontaneous category but display remarkable variation in their subsequent risk of preterm labor and/or infection, which suggests multiple pathways to this still mysterious event (eg, some that progress through short cervix and others that do not).

What does this mean for clinicians who care for women with a previous indicated preterm birth? Should these patients be offered prophylaxis with weekly 17-OHPC? The answer at present is “no,” because supplemental progestogens have not been studied to prevent spontaneous preterm birth in women with a previous indicated preterm birth. Researchers who design future clinical trials of supplemental progestational agents should follow the example of O’Brien et al to record a transvaginal cervical length for all subjects at enrollment. A case can be made currently that any trial of progestogens in women with a previous preterm birth should be powered sufficiently to study women who do and do not have a short cervix. There are 3 trials of progestogens in women with a previous spontaneous preterm birth, only 1 of which measured cervical length. In 2003, da Fonseca et al reported reduced rates of preterm birth in women who were treated with progesterone vaginal suppositories. In that same year, Meis et al found reductions of similar magnitude in women who were treated with 17-OHPC. Neither measured the cervix of the enrolled women. O’Brien et al performed the largest study of women with a previous spontaneous preterm birth. These investigators wisely chose to obtain a cervical length measurement in all enrolled women. The mean cervical length at entry (18-22 weeks of gestation) was 37 mm, and only 4% of enrollees had a cervical length <25 mm. The failure of vaginal progesterone gel to affect the rate of preterm birth in this trial was attributed at first to the formulation, but 2 subsequent successful trials of vaginal progesterone supplementation in women with a short cervix suggest that the results of the trial of O’Brien et al were related to the population (normal cervix), not the preparation of progesterone. One interpretation of currently available information is that a history of spontaneous preterm birth is a clinical marker that is often but not always associated with short cervix. This interpretation of the variable outcomes reported in trials of progestogen supplementation would explain the reductions seen in the Fonseca et al and Hassan et al studies and the absence of benefit in the O’Brien et al and Grobman et al studies. Laughon et al’s results support previous work by Ananth et al showing increased rates of spontaneous preterm birth in women with a prior preterm birth categorized as being indicated. We can conclude that the categorization of a prior preterm birth as spontaneous or indicated does not distinguish populations of women who might benefit from progestogen prophylaxis in future pregnancies. It is reasonable to suggest that women with a prior indicated preterm birth are candidates for a thorough review of the sequence of events leading to that birth and for transvaginal ultrasound screening for short cervix in subsequent pregnancies.