At 12:13 p.m. Pacific Standard Time on September 1, 2017, the San Francisco Bay Area National Weather Service office issued an urgent weather message: “Dangerously hot conditions to begin the Labor Day weekend.” The heat wave set a new record temperature of 106 degrees Fahrenheit in downtown San Francisco, handily beating the previous record of 103 degrees set in 2000.
“Hot temperatures will create a dangerous situation in which heat illnesses are likely,” the message read, advising all San Franciscans to drink plenty of fluids, seek out air-conditioning, and check up on relatives and neighbors. The advisory also warned of heat-related illnesses—particularly for the elderly, children, and sick people—as well as pets and livestock.
Some scientists think another group should be added to the list: pregnant women.
A handful of researchers in the United States, Canada, and elsewhere are methodically accumulating evidence suggesting that higher temperatures could be linked to a higher risk of premature births, stillbirths, or other negative pregnancy outcomes. The findings in each case, while compelling, still raise as many questions as they seem to answer, and all the researchers say that much more work needs to be done. But they also suggest that enough evidence has already surfaced to warrant increased scrutiny—particularly as global warming is expected to drive average temperatures ever upward over coming decades.
“In the future,” said Rupa Basu, chief of air and climate epidemiology at the California Environmental Protection Agency, “this is going to be a growing public-health concern.”
A decade ago, Basu noticed something odd in the scientific literature documenting the health risks of air pollution—a much clearer and well-established relationship. She knew that past research, including some of her own, had shown a link between air pollution and negative pregnancy outcomes, but while the literature alluded to a seasonal pattern, none of the studies controlled for temperature. “I said that some of this must be due to temperature,” Basu recalled, “but we don’t have any data to support that.”
Basu first started to explore the effects of temperature on premature births. Using birth-certificate data from California’s Office of Vital Records, she matched more than 58,000 preterm births occurring during the warm months from 1999 through 2006 with climate data from the state Irrigation-Management Information System and U.S. EPA Air-Quality System. She also pulled air-pollution data from the California Air-Resources Board to assess whether levels of ozone, carbon monoxide, nitrogen dioxide, sulfur dioxide, or smokelike particles were confounding or changing the relationship between temperature and premature births.
For her analysis, Basu used a case-crossover design in which every woman who delivered prematurely served as her own control for comparison. The design eliminates some variables, such as age, that are associated with risk for negative outcomes in pregnancy and could potentially skew the results if different women were compared to one another. She compared temperatures from a few days in the week before the delivery with temperatures on other nearby days, to see if premature births were more likely to happen on or after hotter days.
The results were startling. Her research suggested that an increase of 10 degrees Fahrenheit in weekly average “apparent” temperatures—a combination of heat and humidity—corresponded to an 8.6 percent increase in premature births. That association was independent of air pollution.
Later, she turned her attention to stillbirths, doing a similar temperature analysis with a state registry of fetal death certificates. In March of 2016, Basu published the results from analyzing more than 8,500 stillbirths that occurred during a decade of California’s warm seasons: Stillbirth risk was 10.4 percent higher with a 10-degree Fahrenheit apparent-temperature increase.
After her research on premature birth, the stillbirth results were “pretty much on par with what I was expecting,” Basu said. “I would be shocked if there wasn’t an association.”
These findings have been echoed independently elsewhere. Looking at records of more than 5,000 stillbirths in Quebec over 30 years, Nathalie Auger of Quebec’s institute for public health found that with higher temperatures, stillbirth risk increased continuously for certain categories of stillbirths. For those considered full-term, happening after 37 weeks of pregnancy, the odds of stillbirth were 16 percent higher at 28 degrees Celsius (82 degrees Fahrenheit) compared to 20 degrees Celsius (68 degrees Fahrenheit). The increase in odds of stillbirths between those two temperatures was 19 percent for stillbirths where the cause was marked in the registry as unknown, and 46 percent for those attributed to maternal complications.
Auger and her colleagues hypothesized that higher temperatures could have played a role in those stillbirths with unknown causes, which made up about a quarter of the total. Temperature “is not normally something you would look for” in investigating the cause of a stillbirth to try to prevent a mother from losing another child in the same way, Auger said. “It’s an undiscovered possible cause of stillbirth.”
Pauline Mendola, an epidemiologist at the Eunice Kennedy Shriver National Institute of Child Health and Human Development, along with Sandie Ha, until recently a postdoc at the institute, analyzed medical records on nearly 1,000 stillbirths that occurred between 2002 and 2008 in 15 different U.S. hospital-referral regions from Los Angeles to Miami to Massachusetts. They found that a 1 degree Celsius temperature increase corresponded to a 6 percent increase in stillbirth risk, or about four more stillbirths per 10,000 births.
In addition to their case-crossover study, the group examined the effects of chronic exposure to heat through the whole course of a pregnancy, and were surprised to find the odds of stillbirth were 3.7 times greater when women experienced temperatures that were in the top 10 percent of the range for their location.
“It’s much higher than we would have thought,” Mendola said. “To see something with an odds ratio of three to four—that’s pretty striking.”
Compared to the base rate of stillbirths in the United States—about 24,000 per year in the most recent data—they calculated that the risk increase from heat exposure during pregnancy they observed would translate to about 1,000 additional stillbirths in any given year.
“We were like, ‘wow,’” Ha said. “I think that the prolonged exposure to extreme temperature is actually more important than we thought before.”
The cumulative evidence has been enough for these and other researchers to suggest that previous research on heat vulnerability, which mainly focused on cardiovascular problems in the elderly, didn’t capture the full spectrum of potential threats to public health from rising temperatures.
Pregnant women “have traditionally fallen outside of our conception of who is vulnerable to heat,” said Sabrina McCormick, a sociologist at George Washington University, whose research includes how people respond to climate change—heat in particular. “We need to really change that conception.”
* * *
For all of the compelling research, of course, lots of unanswered questions and important caveats remain. In each study, for example, researchers weren’t looking at the temperatures individual women were experiencing before stillbirths and don’t know how much time women may have spent outside or, more importantly, inside—perhaps with air-conditioning. It would be ideal to have women carry a temperature monitor, said Ha, or assemble a large cohort of women to follow and collect all the potentially interesting variables, said Basu. But such studies would be very expensive to run, and take a long time to get results.
“We’re challenged in our ability to do good work on these questions of rare [health] outcomes and the environment,” Mendola said, because it’s not easy to gather enough cases, with enough detail, to do so. Her study with Ha drew its clinical data on about 1,000 stillbirths from the medical records of nearly 230,000 women giving birth that the National Institute of Child Health and Human Development collected to study risk factors for caesarean deliveries. The birth or death certificates that other studies relied on give only limited information.
Compared to factors like maternal complications, the effect of an environmental exposure on stillbirth risk is small, Ha said, so teasing it out of all the potential confounders is difficult. Some factors that could influence stillbirth risk are closely correlated to temperature, such as air-pollution levels and season of conception, said Tim Bruckner, a public-health researcher at the University of California, Irvine, who has studied the effects of exposure to cold temperatures on birth outcomes in Sweden. “That makes it hard to attribute a causal effect of the birth outcome to temperature.”
Ha and Mendola have also done research on the effects of air pollution on stillbirth, and did control for it as well as season of conception in their temperature study. The effects of air pollution and temperature appear to be independent of one another, Mendola said, “to the extent that the math works.”
But Gary Loy, an obstetrician at Rush University Medical Center in Chicago and advisor to a regional Pediatric Environmental-Health Specialty Unit who was not involved with the temperature- and pregnancy-outcomes research, issued a note of caution. “The strength of association is always in question,” he said, “because there are so many confounders and biases and potential influences.”
Even so, Loy added that, based on what’s been uncovered on temperature and pregnancy thus far, “I think it’s settled there’s an association.” He said, “I don’t think there’s any question.”
Perhaps the biggest caveat is that so far, all the research has been based on observational data. “Epidemiological studies in general have their difficulties,” Loy said. “They’re generally hypothesis-generating studies rather than confirmatory studies.” These studies can show associations but not prove one thing caused the other to happen—a major hurdle for research on the harms of being exposed to various aspects of the environment.
A key question to answer, then, if it’s suspected that outside temperature can impact a child in the womb: What’s the biological explanation for how that could happen?
As of yet, the necessary research to answer that question hasn’t been done, though there are “lots of plausible ties,” Mendola said. Pregnant women, for example, are less able to regulate the temperature of their bodies, which was one reason it made sense to Basu to study the effects of temperature on pregnancy in the first place. Stress from a rising body temperature could also trigger an inflammatory response that constricts a pregnant woman’s blood vessels, making it harder for blood carrying oxygen and other essentials to get to the placenta and putting the baby at risk, Loy said.
The dehydration that accompanies overheating could also play a role, as it decreases the amount of amniotic fluid in the womb, which is associated with fetal death. And there may even be temperature-sensitive proteins in the blood vessels of the placenta and fetus that cause the vessels to get wider, dropping blood pressure and threatening blood supply to the fetus through another theoretical mechanism, said Eric Benner, a neonatologist at Duke University.
McCormick also wonders if there is a window of time within pregnancy when a baby in utero is particularly vulnerable to heat, and if a prolonged exposure to warm temperatures throughout pregnancy—or an extreme but short heat wave—is more hazardous. Does temperature have to rise beyond a threshold? If so, what is it?
Even repeating the same types of studies that have already been done, with new datasets, would be valuable, the researchers suggest. “Really, some of the grunt work of replication is needed right now,” Bruckner said. “It’s not so flashy.”
Basu and others, including Ha, who has taken a new position at the University of California, Merced, see enough intriguing evidence to continue their research. “We’ll be seeing more and more of this evidence,” Basu said. “It’s just not there quite yet.”
Nonetheless, McCormick would like to see pregnant women included in public-health advisories about heat that currently target the elderly. “I do think that we have enough research at this point to be concerned about pregnant women as a vulnerable population,” she said. “We should be warning pregnant women about the risks of heat.” Unlike other sources of risk for stillbirth, heat is something pregnant women can try to avoid or combat, such as by spending time in air-conditioning and staying hydrated, the researchers say.
“It’s pretty much everybody in this population is exposed,” Basu said. “It has the ability to really affect a lot of people.”