Fat pregnant mums and birth complications — study decoding
First-time mums heard worrisome news when The Times reported that a new study had ostensibly found that fat women were more likely to develop preeclampsia during pregnancy and to deliver low birthweight and premature babies. This study, first presented at the 2007 annual meeting of the Society for Gynecological Investigation in Reno, Nevada, also reported that cesarean section rates among nulliparous pregnant fat women were the “highest ever reported in the world.”
Without an understanding of “fair test,” which is how to judge the quality of a study to know if we can rely on it to make any health decisions, readers might never have recognized the major flaws in the design of this research and why this study was unable to support any of the claims that have frightened pregnant women around the world. This news story also illustrated the necessity of going to the original study and reading critically what it’s actually saying, not what the words seem to be saying.
Before we look at the study, first the news. As Health Editor, Sam Lister, reported:
First-time mothers who are obese are almost twice as likely [the study didn’t even look at any women who weren’t obese, all were high risk with PE] to have premature and low birthweight babies and have a higher risk of pre-eclampsia, research suggests. Those with a body mass index (BMI) of more than 30 carry a particularly high risk, according to a study of 385 obese British and Dutch women in their first pregnancy.
Of the babies born to these women, 18.8 per cent were of lower than normal birthweight compared with a rate of around 10 per cent for the general population. A total of 13.4 per cent were higher than normal birthweight; the average is 10 per cent. And 11.7 per cent of the women developed pre-eclampsia, compared with 2 per cent of the average population and 6 per cent of obese women with one or more previous pregnancies. The rate of pre-eclampsia increased as BMI rose, according to the study published in the American Journal of Obstetrics and Gynaecology. The premature birth rate in the group was 11.9 per cent; the national average is 7 per cent.
The Caesarean section rate of 39 per cent was the highest ever reported in the world, the authors said. The higher their BMI, the more mothers were likely to need to stay in hospital. Lucilla Poston, who led the research, funded by the baby charity Tommy's, the Wellcome Trust and the Biomedical Research Centre, at Guy's & St Thomas' NHS Foundation, said …“The large proportion of small babies was particularly unexpected, as obesity is more often associated with the birth of overweight babies.”
This study, led by professor Lucilla Postaon, Ph.D., from Kings College London, and midwife colleagues, opened by saying that “the recent Foresight Report has emphasized the extent of the problem [of obesity] in the UK.”[Referencing the Foresight Tackling Obesities: Future Choices Project report might have been a heads up that we could be in store for something other than the soundest of science. The Foresight report, as readers know, was based on no scientific evidence at all, but an incredible series of assumptions, contradictions and speculations, while proposing some of the most massive governmental anti-obesity policies in the history of the world.]
The aim of this study published in AJOG, wrote professor Poston and colleagues, was to examine the poor pregnancy outcomes and utilization of healthcare resources by fat nulliparous women.
But this study was not a fair test of this objective. The most serious flaw was allocation bias. That’s where the study and control groups are weighted to make obesity appear to be far more unhealthy than it actually is. The group of women chosen to be examined were not taken from the general population, nor was there any randomization to create the study and control groups to enable impartial comparisons.
There was no control group at all.
For this study population, the authors turned to a study they’d published in 2006 in Lancet. That study was a randomized clinical trial of vitamin C and vitamin E supplements in a high-risk population of pregnant women. To enroll in the vitamin trial, the women had to have at least one of eight clinical risk factors for pre-eclampsia. [That clinical trial, by the way, had shown that the 2,404 women taking the anti-oxidant vitamins had a 28% incidence of low birthweight babies, compared to 24% in the control group. The vitamins slightly raised the rate of babies born with low birthweights.]
For this new paper, the authors did a secondary analysis of 385* of the 408 ‘obese’ nulliparous (first pregnancies) women from the placebo arm of this trial. In other words, the authors were only looking at fat women at high risk for pre-eclampsia. Yet, these women were reported as being representative of all fat women and their fat blamed for them being at high risk!
So, the claims that fat first-time moms are twice as likely to have pregnancy complications — ask: “Twice as likely as who?” Compared to other high-risk moms? No. Compared to other first-time moms? No. Their high-risk cohort of first-time mothers was compared to the general population of healthy women of all parities.
While the authors also reported that risks rose with increasing BMI, that is also not what their own data showed. Comparing the women with a BMI of 35 to the most extremely obese, with BMIs over 40, for instance, finds:
● only a 6% higher rate of pre-eclampsia — based on two women
● no difference in spontaneous pre-term births
● fewer (21%) very preterm births, under 34 weeks
● no difference in babies admitted to neonatal intensive care units
● fewer (8%) instrument deliveries
● lower rates of elective cesarean sections
● lower rates of emergency cesarean sections
● a mere 7/10 of a day longer post-natal hospital stay
● fewer extremely low birthweight babies (under 5th percentile)
● fewer low birthweight babies (under 10th percentile) — 18% compared to 24%
● no difference in all low birthweight babies under 2,500 grams
Clearly, rising BMIs were not associated with rising risks for these mothers or for their babies.
Low birthweight. First, the evidence failed to support the authors’ statement that fat nulliparous women were of greater risk for delivering underweight (growth-restricted) babies. Even they noted that this finding had never been recognized in the medical literature before and that it contradicted the body of medical evidence, which has found obesity to be protective against fetal growth retardation.
Among the most recognized factors in underweight babies are severe stress, associated with two-fold increase in low birthweight babies and premature babies, thin mothers, poor pregnancy weight gain, and various illnesses in mothers.
Pre-eclampsia. The authors also said that “the incidence of pre-eclampsia was greater than generally reported in women of mixed parity.” This was a misleading comparison because all first- time mothers have greater risks for pre-eclampsia compared to other pregnant women. Why? It is well recognized, as the Merck Manual explains, that pre-eclampsia is most common among women who are pregnant for the first time.
Despite the media stories, the authors also reported that the association between BMI and pre-eclampsia was not even statistically significant. But these nuances never made it to the news. Instead, women were left to believe that being fat risked pre-eclampsia.
Pre-eclampsia is also more common among young teen mothers, older women, those carrying more than one baby, and women facing very stressful situations, such as lower class or economic status. The most morbidly obese women also suffer the highest levels of stigma and socio-economic discrimination. Yet, none of these important confounding factors were considered when this cohort was compared with the general population.
Preemies. The authors reported that preterm births among these fat first-time mothers were nearly twice the incidence of women in the general population. These rates have never been noted to the degree reported in this study, they wrote.
Here, again, their own data failed to show that rising BMI even correlated with more premature babies. They also compared this high risk population to the general population. But, as reported earlier this year, doctors with the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) noted that it’s been long documented that the risk for spontaneous preterm births decrease with increasing maternal BMI.
Cesarean sections. The authors wrote that cesarean section rates among the ‘obese’ first-time mothers “were higher than those reported in any study of obese pregnancy in any population worldwide.” But, there was no correlation (nor even a statistical association) between the mothers’ weights and C-section rates, and even the authors noted that another study of women in the UK found the incidences of C-sections were less than half those in this study. Here again, complications among a high-risk population can’t credibly be applied to all fat women.
Hospital stays. Finally, the authors compared inpatient hospital stays among these high-risk women to normal weight women in the general population, not among women of similar high risk, making it not viable to attribute any differences to BMI. The authors supported their assertion that fat women have longer hospital stays, however, by citing the Kaiser Permanente study, covered here. Examining nearly 20,000 consecutive births, Kaiser researchers had failed to find any real difference — reporting a mere 0.1 day difference — between women of “normal” weight and the most extremely “obese.”
There were other limitations in this study, such as the gestation of the women at enrollment had been poorly controlled, meaning women further along in gestation would more likely have had higher BMIs and to have been misclassified. The “routine collection of data was sporadic and incomplete among the trial centers,” they wrote. But the most important limitation was its failure to have a control group, drawn from the same population, for comparison.
Scientists would have designed a very different study: a fair test to objectively explore whether or not obesity was associated with higher risks. This study was not that.
A study that is not a fair test is not a study we can use. It may make good fodder for media, but not for basing medical decisions. This study failed to provide sound evidence to support any of the claims heard in the news. Nor could the authors give a biological explanation for why fat itself could give these findings. And finally, the body of medical research to date contradicts the heightened obesity scares.
The study’s conclusion is a final test of reading comprehension. The statements are factual, but what they say is quite different from what the casual reader might easily believe they are saying:
In conclusion, this prospective study of a contemporary cohort highlights the need to address parity when assessing risk among obese pregnant women. The high incidences of SGA [small for gestational age] and preterm birth have clinically important implications for serial fetal surveillance, and both will contribute to the increased use of health care resources.
Prenatal care is important for all first-time mothers. And underweight and premature babies have important implications for their health and healthcare costs. But that doesn’t mean this study demonstrated the mothers’ obesity to be the reason.
© 2009 Sandy Szwarc
* No mention was made of why 23 were excluded.