Metformin Use in Nondiabetic Obese Pregnancy

Article from The Daily Mail, 2011

One of the strongest concerns doctors have about pregnancies in the "obese" is that larger people tend to have larger (macrosomic) babies. Although most macrosomic babies are born just fine, they do have higher rates of shoulder dystocia (babies who get stuck) and related injuries, as well as low blood sugar at birth and more cesareans. So doctors want to do everything they can to prevent abnormally big babies.

Some macrosomia is tied to high blood sugar and high insulin levels. So in hopes of preventing big babies, doctors have been using the diabetes medication, metformin, in those diagnosed with Gestational Diabetes (GD) or Polcystic Ovarian Syndrome (PCOS).

A number of studies have confirmed that metformin use in women with GD does modestly reduce the rate of big babies. It also lowers the rate of early pregnancy loss and prematurity in PCOS. More research is needed but metformin does seem to be a very helpful drug for people with GD or PCOS. No one is questioning this use of metformin.

However, the use of metformin in obese women WITHOUT gestational diabetes or PCOS is a different story. Doctors note that even high BMI people who are not diabetic have larger babies on average. So the working theory has been that these women must be pre-diabetic or have strong insulin resistance that increases fetal size.

So doctors began prescribing metformin to nondiabetic obese women in hopes that lowering insulin levels and borderline blood sugar would cut the odds of a big baby.

The practice was aggressively marketed to the public as a way to prevent "obese babies" before its research was even completed (see headlines quoted here from The Daily Mail 2011 and 2012).

But what does the research say about this use of metformin? Here is a quick summary of the three largest trials.

The Studies on Non-Diabetic High BMI Women

From article in the Daily Mail, 2012

Chiswick 2015

Several years ago, a large study called the EMPOWaR trial (Chiswick 2015) tested this theory in the U.K.

This study involved 15 hospitals and was a large, randomized, double-blind placebo-controlled trial, the gold standard of research. It had n=434 participants with a BMI over 30 for analysis. The maximum metformin dose was 2500 mg.

To authors' great surprise, they found that metformin did NOT lower neonatal size.

Syngelaki 2016

Some common criticisms of the EMPOWaR study were that the metformin dose was too low, the participants weren't fat enough to show any big effect, and they did not take doses strictly enough.

Therefore, in a subsequent study published in the prestigious New England Journal of Medicine (Syngelaki 2016, the MOP trial), n=400 participants were limited to those with a BMI over 35. This study, too, was a randomized, double-blind study with placebo controls and was more racially diverse.

The researchers increased the metformin dose to a maximum of 3000 mg and made sure there was strong adherence to the medication. By limiting the analysis to those with a BMI over 35, increasing the dosage, including more women of color, and making sure metformin was consistently used, the authors hoped to show more of an effect.

To their surprise, results were again similar. While the metformin group had a slightly lower weight gain, fetal size was the same between groups.

Dodd 2019

Researchers just can't leave this theory alone.

Now there is a new study (the GRoW trial) out, also testing the metformin theory (Dodd 2019). This trial was done in Australia and included women with a BMI over 25 (in other words, both "overweight" and "obese"). No previous study had included those in the overweight category.

This also was a gold standard randomized study, n=514 participants. It used doses of up to 2000 mg.

It also found slightly less weight gain in the metformin group but NO difference in birthweight of the babies.

Research Summary

There have been a few other, small studies about metformin use in nondiabetic women, but none have been as large or as strong as these studies. No study so far has found that metformin lowers neonatal birthweight in nondiabetic women. That message is very clear and consistent.

There were other outcomes that weren't as clear. Some, but not all, studies found a mild lowering of prenatal weight gain. Some found decreased incidence of preeclampsia, while others did not. No other outcomes were routinely affected.

At this point, the hypothesis that metformin will "normalize" the size of high BMI women's babies has pretty well been disproven. I'm sure there will be more studies on it because the theory is a favorite of many OBs, but these are strong studies and frankly, I doubt they'll be overturned.

The good news is that no babies seem to have been harmed in these studies. However, many of the mothers experienced significant gastrointestinal side effects from the metformin and this some caused drop-outs or scaled-back dosing. If you've ever taken metformin, you know the G.I. effects can be considerable. This certainly affects people's quality of life. As a result, it's not something that should be prescribed lightly.

The take-home message from research: Metformin is a great drug that can be useful for some indications (like GD or PCOS) but in nondiabetic high BMI women it does not lower neonatal birthweight. As the authors of the EMPOWaR study concluded:

... metformin should not be used to improve pregnancy outcomes in obese women without diabetes.

The Fat-Shaming Around These Studies

Illustration from the 2012 Daily Mail article

It has to be pointed out that the U.K. public health campaign around these studies was glaringly fat-shaming.

Look at the caption above. Fat women are accused of letting their babies be "born obese," of passing on their toxic obesity in the womb through their carelessness about their health. They use the classic picture of a fat body with the head cut off, depersonalizing the subject. The person is even holding a roll of fat, pointing out visual blame so the negative message is even clearer. 

The articles were filled with scary summaries of the risks of obesity and pregnancy, without any context for those risks, how often they don't happen, and what can be done about them when they do. It's not unreasonable to inform women of size of the possible risks around weight and pregnancy, but it's another thing to misrepresent those risks to scare or shame women out of pregnancy.

The campaign was attempting to inflame the public about irresponsible fat people, implying that they refuse to be healthy and are costing the NHS huge amounts of money, taking money away from everyone else. The U.K. is a very fat-phobic place and the government is scapegoating fat people for their healthcare budget woes.

The language of the campaign was also offensive. They used the terms "fat babies" or "obese babies" in order to shame the mothers, but a big baby is not necessarily the same as an "obese" baby. They are conflating fetal size caused by diabetic complications with big babies that are simply larger than average.

All big babies are not alike. Some babies are big because of blood sugar issues, and these babies do tend to be abnormally proportioned and have more issues at birth. On the other hand, some babies are just naturally larger without it being pathological. There is a significant difference between a diabetic's baby that is 9 lbs. but only 16 inches long and a 9 lb. baby that is 22 inches long. The first is abnormal and a true concern; the second is proportional and most likely genetic. The first type often has problems being born safely and has many complications; the second type of big baby is proportional and can usually be born vaginally.

Furthermore, the campaign is simplistic and misleading. Not all obese mothers have macrosomic babies; one study found that only 17% of obese women had macrosomic babies while 83% of them did NOT. Subjecting all obese women to metformin "just in case" means medicating many people who wouldn't produce a big baby anyhow. What potential harm might that be doing?

Some people of average size also have macrosomic babies without blood sugar or insulin issues; no one knows why some babies are bigger than others. And many big babies do have vaginal births; Navti 2007 found that 83% of women who had babies around 10 pounds or more were able to have vaginal births. This shows that even very big babies can often be born vaginally, given time, patience, sufficient mobility, and a calm caregiver. We need to stop panicking over babies that are larger than average and save our intervention for those who truly need it.

Researchers: Stop trying to put the baby on a diet before it is even born. Metformin for reducing fetal size does not work in nondiabetics. 

Public Health Campaigns: Stop promoting weight stigma and fat-shaming in your campaigns about obesity and pregnancy. 



References

Lancet Diabetes Endocrinol. 2019 Jan;7(1):15-24. doi: 10.1016/S2213-8587(18)30310-3. Epub 2018 Dec 4. Effect of metformin in addition to dietary and lifestyle advice for pregnant women who are overweight or obese: the GRoW randomised, double-blind, placebo-controlled trial. Dodd JM, Louise J, Deussen AR, Grivell RM, Dekker G, McPhee AJ, Hague W.  PMID: 30528218

... GRoW was a multicentre, randomised, double-blind, placebo-controlled trial in which pregnant women at 10-20 weeks' gestation with a BMI of 25 kg/m2 or higher were recruited from three public maternity units in Adelaide, SA, Australia. Women were randomly assigned (1:1) via a computer-generated schedule to receive either metformin (to a maximum dose of 2000 mg per day) or matching placebo. Participants, their antenatal care providers, and research staff (including outcome assessors) were masked to treatment allocation...  FINDINGS: Of 524 women who were randomly assigned between May, 28 2013 and April 26, 2016, 514 were included in outcome analyses (256 in the metformin group and 258 in the placebo group). Median gestational age at trial entry was 16·29 weeks (IQR 14·43-18·00) and median BMI was 32·32 kg/m2 (28·90-37·10); 167 (32%) participants were overweight and 347 (68%) were obese. There was no significant difference in the proportion of infants with birthweight greater than 4000 g (40 [16%] with metformin vs 37 [14%] with placebo; adjusted risk ratio [aRR] 0·97, 95% CI 0·65 to 1·47; p=0·899). Women receiving metformin had lower average weekly gestational weight gain (adjusted mean difference -0·08 kg, 95% CI -0·14 to -0·02; p=0·007) and were more likely to have gestational weight gain below recommendations (aRR 1·46, 95% CI 1·10 to 1·94; p=0·008). ... INTERPRETATION: For pregnant women who are overweight or obese, metformin given in addition to dietary and lifestyle advice initiated at 10-20 weeks' gestation does not improve pregnancy and birth outcomes.

N Engl J Med. 2016 Feb 4;374(5):434-43.doi: 10.1056/NEJMoa1509819. Metformin versus Placebo in Obese Pregnant Women without Diabetes Mellitus. Syngelaki A, Nicolaides KH, Balani J, Hyer S, Akolekar R, Kotecha R, Pastides A, Shehata H. PMID: 26840133

[kmom summary] Randomized double-blind, placebo controlled trial. Limited to those with BMI over 35 and upped the metformin dosage. Less preeclampsia and less weight gain in metformin group but no difference in birth weight. "CONCLUSIONS: Among women without diabetes who had a BMI of more than 35, the antenatal administration of metformin reduced maternal weight gain but not neonatal birth weight."

Lancet Diabetes Endocrinol. 2015 Oct;3(10):778-86. doi: 10.1016/S2213-8587(15)00219-3. Epub 2015 Jul 9. Effect of metformin on maternal and fetal outcomes in obese pregnant women (EMPOWaR): a randomised, double-blind, placebo-controlled trial. Chiswick C, Reynolds RM, Denison F, Drake AJ, Forbes S, Newby DE, Walker BR, Quenby S, Wray S, Weeks A, Lashen H, Rodriguez A, Murray G, Whyte S, Norman JE. PMID: 26165398 Free full text here.

[kmom summary] Randomized placebo-controlled, double-blind study in 15 hospitals in the U.K. on nondiabetic women. Results: "Metformin has no significant effect on birthweight percentile in obese pregnant women."

Previous discussion of these studies and others:

• https://wellroundedmama.blogspot.com/2011/12/fourth-annual-turkey-awards-leaps-of.html
• https://wellroundedmama.blogspot.com/2015/10/metformin-does-not-lower-birthweight.html
• https://wellroundedmama.blogspot.com/2016/03/once-again-metformin-does-not-reduce.html

Metformin for Gestational Diabetes or PCOS

J Matern Fetal Neonatal Med. 2018 Nov 20:1-141. doi: 10.1080/14767058.2018.1550480. [Epub ahead of print] Metformin-treated-GDM has lower risk of macrosomia compared to diet-treated GDM- A retrospective cohort study. Bashir M, Aboulfotouh M, Dabbous Z, Mokhtar M, Siddique M, Wahba R, Ibrahim A, Al-Houda Brich S, Konje JC, Abou-Samra AB. PMID: 30458653

...This is a retrospective cohort study that included GDM women compared to normoglycaemic controls between March 2015-December 2016 in the Women's Hospital, Qatar. RESULTS: The study included 2221 women; of which 1420 were normoglycaemic, and 801 were GDM (358 GDM-D and 443 GDM-T)... Women in the GDM-T group had lower GWG/week compared to GDM-D (-0.01 ± 0.7 versus 0.21 ± 0.51 kg/week; p < 0.001). After correcting for age, prepregnancy weight and GWG; GDM-T had higher risk of preterm labour (OR 1.66; 95% CI 1.20-2.22), and C-section (OR 1.37, 95% CI 1.02-1.85) and reduced risk of macrosomia (OR 0.56; 95% CI 0.32-0.96) and neonatal hypoglycaemia (OR 0.49; 95% CI 0.28-0.82). CONCLUSION: ... Treatment with metformin reduces maternal weight gain, the risk of macrosomia and neonatal hypoglycaemia compared to diet alone.

J Clin Endocrinol Metab. 2010 Dec;95(12):E448-55. doi: 10.1210/jc.2010-0853. Epub 2010 Oct 6. Metformin versus placebo from first trimester to delivery in polycystic ovary syndrome: a randomized, controlled multicenter study. Vanky E et al.  PMID: 20926533

[kmom summary] n=274 PCOS pregnancies. Randomized controlled trial with placebos. Less prematurity, but more pre-eclampsia in metformin group. Less weight gain in metformin group. No difference in fetal size between groups.  

Estimating Fetal Weight Increases Risk for Cesarean

Think twice about doing ultrasounds to estimate fetal weight before birth.

In this very large, multi-center study, just the act of estimating fetal weight raised the cesarean rate if the baby was predicted to be big, even when controlling for actual fetal size. It doubled the risk for cesarean in non-diabetic mothers who were thought to be carrying large babies.

Most women predicted to have a large baby will not actually have a large baby, yet fear of a large baby lowers the surgical threshold for many providers, resulting in unnecessary cesareans. Despite limited evidence of improved outcomes, estimating fetal weight is a very common intervention in most obstetric practices, particularly for women of size who tend to have larger babies on average. It is likely a major driver of the high cesarean rate in "obese" women.

Providers need to stop doing so many fetal weight estimates and over-managing the labors of suspected big babies. This is especially important in women of size.


Reference

Obstet Gynecol. 2016 Sep;128(3):487-94. doi: 10.1097/AOG.0000000000001571. Association of Recorded Estimated Fetal Weight and Cesarean Delivery in Attempted Vaginal Delivery at Term. Froehlich RJ1, Sandoval G, Bailit JL, Grobman WA, Reddy UM, Wapner RJ, Varner MW, Thorp JM Jr, Prasad M, Tita AT, Saade G, Sorokin Y, Blackwell SC, Tolosa JE; MSCE, for the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Maternal-Fetal Medicine Units (MFMU) Network. PMID: 27500344

OBJECTIVE: To evaluate the association between documentation of estimated fetal weight, and its value, with cesarean delivery. METHODS: This was a secondary analysis of a multicenter observational cohort of 115,502 deliveries from 2008 to 2011. Data were abstracted by trained and certified study personnel. We included women at 37 weeks of gestation or greater attempting vaginal delivery with live, nonanomalous, singleton, vertex fetuses and no history of cesarean delivery. Rates and odds ratios (ORs) were calculated for women with ultrasonography or clinical estimated fetal weight compared with women without documentation of estimated fetal weight. Further subgroup analyses were performed for estimated fetal weight categories (less than 3,500, 3,500-3,999, and 4,000 g or greater) stratified by diabetic status. Multivariable analyses were performed to adjust for important potential confounding variables. RESULTS: We included 64,030 women. Cesarean delivery rates were 18.5% in the ultrasound estimated fetal weight group, 13.4% in the clinical estimated fetal weight group, and 11.7% in the no documented estimated fetal weight group (P<.001). After adjustment (including for birth weight), the adjusted OR of cesarean delivery was 1.44 (95% confidence interval [CI] 1.31-1.58, P<.001) for women with ultrasound estimated fetal weight and 1.08 for clinical estimated fetal weight (95% CI 1.01-1.15, P=.017) compared with women with no documented estimated fetal weight (referent). The highest estimates of fetal weight conveyed the greatest odds of cesarean delivery. When ultrasound estimated fetal weight was 4,000 g or greater, the adjusted OR was 2.15 (95% CI 1.55-2.98, P<.001) in women without diabetes and 9.00 (95% CI 3.65-22.17, P<.001) in women with diabetes compared to those with estimated fetal weight less than 3,500 g. CONCLUSION: In this contemporary cohort of women attempting vaginal delivery at term, documentation of estimated fetal weight (obtained clinically or, particularly, by ultrasonography) was associated with increased odds of cesarean delivery. This relationship was strongest at higher fetal weight estimates, even after controlling for the effects of birth weight and other factors associated with increased cesarean delivery risk.

Estimating Fetal Weight Increases Cesarean Risk

Image from Wikimedia

One intervention often seen in pregnancies these days is estimation of fetal weight. But does this improve outcomes?

Bigger babies are more at risk for getting stuck, which is called "shoulder dystocia." Although the actual numerical risk for shoulder dystocia with big babies is not very great (and small babies get stuck too), the risk for shoulder dystocia is higher with a bigger baby. Most of the time it is resolved without damage, but in some cases babies or moms are hurt, sometimes seriously. In rare cases, babies even die. Damage from shoulder dystocia is one of the main reasons care providers get sued.

As a result, a big baby make many care providers nervous. So they routinely employ estimation of fetal weight near the end of pregnancy, either by hands-on estimation or by ultrasound estimation. But the critical question is whether estimating fetal weight actually improves outcomes.

Here is yet another study that shows that estimating fetal weight does NOT improve outcomes and that it increases the risk for cesareans. 

In this very large multi-center MFMU study, a baby estimated to be above 4000g (8 lbs., 13 oz.) was more than twice as likely to be born by cesarean. A diabetic mother with a baby estimated to be above 3500g (about 7 lbs. 12 oz.) was more than nine times as likely to have a cesarean.

This relationship persisted even when adjusted for actual birth weight and other variables, showing it was not just about the baby's size. The mere prediction of a large baby (whether baby was actually large or not) increased the risk for cesarean, suggesting that it is the way big babies are managed that makes the difference.

This is a particularly pertinent issue for women of size. Because "obese" women tend to have larger babies on average, care providers can get very anxious about fetal size in their pregnancies. Fear of big baby (and the resulting interventions) is one of the top drivers around the outrageously high cesarean rate in high-BMI women.

Sadly, this study does not directly address the effect of fetal weight estimation in the sub-population of obese women. It would be great to have data actually examining a practice that is so common in the management of obese women.

However, this is a very large and very powerful study showing that the practice of estimating fetal weight is harmful in pregnant women in general.

And it is just the latest in a long series of studies that show that the common practice of estimating fetal weight increases the cesarean rate.

Given that it is such a common practice in obstetrics these days, the practice of estimating fetal size before birth needs to be re-evaluated. 

There are times when estimating fetal size can be appropriate, but most of the time it is more harmful than helpful. Care providers need to stop using it routinely.

And its use especially needs to be evaluated in the management of women of size.


Reference

Obstet Gynecol. 2016 Sep;128(3):487-94. doi: 10.1097/AOG.0000000000001571. Association of Recorded Estimated Fetal Weight and Cesarean Delivery in Attempted Vaginal Delivery at Term. Froehlich RJ1, Sandoval G, Bailit JL, Grobman WA, Reddy UM, Wapner RJ, Varner MW, Thorp JM Jr, Prasad M, Tita AT, Saade G, Sorokin Y,Blackwell SC, Tolosa JE; MSCE, for the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Maternal-Fetal Medicine Units (MFMU) Network. PMID: 27500344

OBJECTIVE: To evaluate the association between documentation of estimated fetal weight, and its value, with cesarean delivery. METHODS: This was a secondary analysis of a multicenter observational cohort of 115,502 deliveries from 2008 to 2011. Data were abstracted by trained and certified study personnel. We included women at 37 weeks of gestation or greater attempting vaginal delivery with live, nonanomalous, singleton, vertex fetuses and no history of cesarean delivery. Rates and odds ratios (ORs) were calculated for women with ultrasonography or clinical estimated fetal weight compared with women without documentation of estimated fetal weight. Further subgroup analyses were performed for estimated fetal weight categories (less than 3,500, 3,500-3,999, and 4,000 g or greater) stratified by diabetic status. Multivariable analyses were performed to adjust for important potential confounding variables. RESULTS: We included 64,030 women. Cesarean delivery rates were 18.5% in the ultrasound estimated fetal weight group, 13.4% in the clinical estimated fetal weight group, and 11.7% in the no documented estimated fetal weight group (P<.001). After adjustment (including for birth weight), the adjusted OR of cesarean delivery was 1.44 (95% confidence interval [CI] 1.31-1.58, P<.001) for women with ultrasound estimated fetal weight and 1.08 for clinical estimated fetal weight (95% CI 1.01-1.15, P=.017) compared with women with no documented estimated fetal weight (referent). The highest estimates of fetal weight conveyed the greatest odds of cesarean delivery. When ultrasound estimated fetal weight was 4,000 g or greater, the adjusted OR was 2.15 (95% CI 1.55-2.98, P<.001) in women without diabetes and 9.00 (95% CI 3.65-22.17, P<.001) in women with diabetes compared to those with estimated fetal weight less than 3,500 g. CONCLUSION: In this contemporary cohort of women attempting vaginal delivery at term, documentation of estimated fetal weight (obtained clinically or, particularly, by ultrasonography) was associated with increased odds of cesarean delivery. This relationship was strongest at higher fetal weight estimates, even after controlling for the effects of birth weight and other factors associated with increased cesarean delivery risk.

Most Babies Suspected to be Large Actually Aren't

So many mothers get the "big baby" fear card handed to them during pregnancy. This often leads to pressure for interventions such as induction of labor, or scaring women into a planned cesarean.

In this survey, one-third of women surveyed were told their babies were getting "quite large." Yet only one in five of these women actually went on to have large babies (around 9 lbs. or more).

Although the survey doesn't have information specifically about women of size, I would bet that the rate of big baby fear cards is even higher in our care. There is some justification for this; on average, we do tend to have somewhat larger babies. But even so, most high-BMI women have average-sized babies. And even if they do have a bigger baby, most of these can be born just fine with a little care and patience, especially if they have full mobility during labor.

But the fear of big babies from big mothers is so exaggerated these days that many care providers act as if every high-BMI woman will have a Godzilla Baby. Sadly, a great deal of the increased interventions we experience is likely tied to that fear of big babies.


Reference

Matern Child Health J. 2015 Dec;19(12):2578-86. doi: 10.1007/s10995-015-1776-0. Labor and Delivery Experiences of Mothers with Suspected Large Babies. Cheng ER1, Declercq ER2, Belanoff C3, Stotland NE4, Iverson RE5. PMID: 26140835

OBJECTIVE: To characterize the prevalence of and factors associated with clinicians' prenatal suspicion of a large baby; and to determine whether communicating fetal size concerns to patients was associated with labor and delivery interventions and outcomes. METHODS: We examined data from women without a prior cesarean who responded to Listening to Mothers III, a nationally representative survey of women who had given birth between July 2011 and June 2012 (n = 1960). We estimated the effect of having a suspected large baby (SLB) on the odds of six labor and delivery outcomes. RESULTS: Nearly one-third (31.2%) of women were told by their maternity care providers that their babies might be getting "quite large"; however, only 9.9% delivered a baby weighing ≥4000 g (19.7% among mothers with SLBs, 5.5% without). Women with SLBs had increased adjusted odds of medically-induced labor (AOR 1.9; 95% CI 1.4-2.6), attempted self-induced labor (AOR 1.9; 95% CI 1.4-2.7), and use of epidural analgesics (AOR 2.0; 95% CI 1.4-2.9). No differences were noted for overall cesarean rates, although women with SLBs were more likely to ask for (AOR 4.6; 95% CI 2.8-7.6) and have planned (AOR 1.8; 95% CI 1.0-4.5) cesarean deliveries. These associations were not affected by adjustment for gestational age and birthweight. CONCLUSIONS FOR PRACTICE: Only one in five US women who were told that their babies might be getting quite large actually delivered infants weighing ≥4000 g. However, the suspicion of a large baby was associated with an increase in perinatal interventions, regardless of actual fetal size.

Metformin Does NOT Lower Birthweight Among Non-Diabetic Obese Mothers

Although most higher-BMI women have average-sized babies, larger women have a higher percentage of big babies. This is one of the biggest reasons care providers intervene in the pregnancies of women of size.

Although most big babies do just fine, bigger babies do have higher rates of issues like shoulder dystocia, cesarean birth, and low blood sugar after birth. So care providers have long searched for ways to lower the rate of big babies among women of size.

Whether that's justified or not is a debate for another day. The point is that many care providers are willing to go to extreme lengths for this goal.

A few years ago there was a large public campaign pushing the prescription of metformin (brand name: Glucophage) for reducing birthweights among non-diabetic "obese" mothers.

We've written about this before. The study was called EMPOWaR and was a randomized, controlled study at 15 different U.K. hospitals.

The theory was that insulin resistance and/or borderline blood sugars were probably at the root of a higher incidence of large infants among high-BMI women, and that lowering blood sugar and insulin resistance even in non-diabetic obese mothers might improve outcomes.

While the publicity campaign noted that they were just investigating this possibility ─ really! ─ the publicity push around a study that hadn't even been done yet suggests that the investigators really had ulterior motives.

Pushing Unproven Agendas Through Publicity 

This is one of my pet peeves about research on obesity in pregnancy; it is often publicized now before the study is even done. One suspects that the researchers are trying to promote unproven high-intervention protocols for this group, trying to raise their own public profiles, or maybe even create a new market for certain medications or programs.

Why else would researchers publicize a study not yet even done?

Publicizing a research trial before it's done creates an expectation in the reading public, including other doctors, that a particular protocol or medication is THE way to manage a particular population or problem. It does an end run around the usual research procedures and starts promoting protocol changes in the minds of the public without having to wait for any pesky results.

We've seen it before in a Kaiser study that promoted zero weight gain in obese pregnant women before the study had even been done. Best guess is that it was part of a push from some doctors to lower prenatal weight gain guidelines because they didn't think the 2009 guidelines (11-20 lbs. for obese women) were low enough.

Certainly, doctors can campaign for changes to national guidelines because of strongly-held beliefs that a particular approach might improve outcomes. However, a pet theory is not enough; any changes need to be supported by actual evidence.

It is medically unethical for doctors to be promoting changes to policy without having clear data that shows a need for such changes and conclusive proof of a lack of harm of such changes.

And guess what? There is good reason for demanding such proof of lack of harm before guideline changes. Turns out that there IS harm in minuscule weight gains.

A recent meta-analysis of 18 cohort studies concluded that "gestational weight gain below the guidelines cannot be routinely recommended" in obese women because of an increase in too-small infants and premature births.

THAT is why you need to show proof of a lack of harm of proposed changes and why no one should be publicizing studies before they are done. Doctors all over the country ─ outside of the research trials ─ took that publicity push seriously and are currently promoting minuscule gains, zero gain, and even weight loss in obese pregnant women. How many premature babies and too-small babies have been the result?

Promote healthy eating and lifestyle? Absolutely. And some larger women naturally gain very little in pregnancy, even with normal eating. That's okay; those women generally do okay. But to deliberately manipulate the diet so that weight gains are minuscule or non-existent? Unwise.

Too many care providers have jumped on the minuscule weight gain bandwagon because of the publicity push promoting it before these protocols were examined adequately for harm.

The Metformin Study

A similar questionable publicity campaign surrounded the British study using metformin to hopefully lower birth weights in babies of obese women. Some publicity articles used scare tactics and hyperbole about pregnancy risks in obese women to justify using this medication experimentally and implied that taking metformin would keep "overweight" women from having "overweight" babies. Basically, it was implied that obese women would be irresponsible if they didn't comply with this experiment.

The study had other issues; it used birthweight as a surrogate marker for the future ill-health of obese mothers' offspring. To me this is a very questionable assumption, since it is difficult to untangle cause and effect of metabolic issues like PCOS, lipedema, thyroid disturbances, and genetic contributions to a child's health vs. birthweight alone. In addition, big babies don't automatically become unhealthy adults, and there is plenty of evidence that too-small babies have more health problems than bigger ones.

The publicity campaign fanned the anti-obesity hysteria in the U.K., created a climate where women in the study probably felt they had little choice about taking such drugs (even though their use in this context was experimental), and created an expectation among care providers that metformin was the standard of care even in the pregnancies of non-diabetic obese women.

So I didn't love the premise of the study or the means by which they pressured women into it, even though I thought it would be interesting to see if metformin had an effect on birthweight.

Well, the study results are in, and metformin did NOT reduce birthweight among obese mothers. Birthweights were similar between the metformin and placebo groups. The ponderal index (a measure of length and weight, kind of like a BMI for babies) was similar between groups as well. In other words, metformin not only didn't lower average birthweight, it didn't make the babies any skinnier for their lengths either.

Nor did metformin improve other outcomes of pregnancy in obese women. The metformin group did not have lower prenatal weight gains, nor did they have fewer cesareans.

Basically, the researchers could not show any meaningful improvements in outcome from taking metformin.

There were a few minor differences between groups that did not rise to statistical significance. The metformin group did develop fewer cases of gestational diabetes, but only marginally. There were some small improvements in blood sugar and insulin in the metformin group at 28 weeks, but not at 36 weeks. Some inflammatory markers were lessened but the significance of this is unclear and didn't seem to have any bearing on immediate outcomes.

The metformin group did develop more pregnancy-induced hypertension, pre-eclampsia, and pre-term birth, but again only marginally. The difference did not rise to statistical significance.

Interestingly, there was a stronger tendency towards more poor outcomes (miscarriage, termination, stillbirth, or neonatal death) among the metformin group (3% vs. 1%), but the confidence interval crossed 1.0 and these results could have been mere coincidence rather than a real result of metformin. Given the rarity of such poor outcomes, the study group probably was not large enough to determine whether a real relationship between metformin and poor outcomes in non-diabetic mothers exists. In the study's defense, the description of the poor outcomes does not seem to indicate that they were due to metformin. The bottom line is that the study did not find a statistically significant increase in poor outcomes among those taking metformin.

It IS important to note that metformin is probably a relatively safe drug to use in pregnancy and is used with significant benefit in some diabetic mothers and women with PCOS, but its safety and efficacy in other mothers is less established.

Now we know that it really doesn't lower the birthweight of babies nor improve other outcomes in non-diabetic high BMI mothers. That led the study authors to conclude:

Metformin should not be used to improve pregnancy outcomes in obese women without diabetes.

Further Details

The fact that the study found no real benefit from metformin use in non-diabetic obese women disproved the authors' hypothesis that metformin would improve outcomes. However, the study's authors speculate that an impact on birthweight was not seen because of several possibilities.

First, the medication was not started until 12-16 weeks, rather than early in pregnancy or pre-conception. They theorized that perhaps the programming of fetal size takes place so early that starting it at 12-16 weeks was too late. However, since most women aren't seen in pregnancy until the end of the first trimester, it is unlikely most obese pregnant women would be able to be started earlier anyhow.

Second, they wondered if the doses may not have been high enough to be effective. They started at 500 mg per day and slowly increased the dosage until the "maximum tolerable dose" was reached ─ i.e., until woman experienced too many side effects like nausea, vomiting, diarrhea to tolerate continuing to increase the dose. Still, about 2/3 of the metformin arm received 2000 mg, very near the maximum dosage of 2500 mg, so the argument that the dose wasn't high enough is weak. This was an adequate test.

Third, the authors speculate that the real benefits of taking metformin during the fetal period would likely show in other ways as the child grew up instead of affecting birthweight. They cite an animal study that suggests less visceral fat in the offspring of mothers that received metformin during gestation. As a result, they are planning a follow-up study to monitor metformin-exposed children and see if they have less obesity and/or metabolic issues as they grow up.

This last theory is one that a lot of doctors are fixating on. This is the idea of fetal programming, that the window of time during gestation is one in which fat women "program" their babies to be larger and to have poorer long-term health, and therefore we needed to be extremely proactive about intervening in the pregnancies of obese women. One commentator noted:

The bold idea that what we do to the fetus during the short and finite period of pregnancy could change and even improve lifelong outcomes of offspring validates the whole concept of prenatal care. If this concept is true, this tiny window of opportunity should not be wasted.

This is an alarming statement to me. Yes, if we could change things during the fetal period that would improve that child's health long-term, that would be an exciting possibility. However, the potential for abuse here is quite high. I worry that researchers are SO excited about "preventing" obesity that their common sense will go out the window and they will start using even more scorched-earth ─ and unproven ─ tactics.

What if we intervene and change the child's long-term health negatively? I worry that scorched-earth protocols to get smaller babies through gestational weight loss or medications may actually backfire and create ripples those care providers don't anticipate. Where are the safety protocols to ensure lack of harm from such interventions?

The in-utero time is a powerful time, it's true. But that means we must be very VERY careful in how we intervene, if we intervene at all. And we certainly shouldn't be publicizing a particular approach until it has been proven both beneficial and harmless.

Conclusion

Big mothers tend to have bigger babies on average. This leads many care providers to institute major interventions and scorched-earth protocols to lower birth weight.

The most common intervention is to limit prenatal weight gain. While women of size probably need to gain less weight in pregnancy than other women, how much weight they should (or shouldn't) gain is more controversial. Even more controversial is what should be done to try to achieve that lower gain.

Too much weight gain is clearly linked to larger babies, and perhaps to other poor outcomes. As a result, many care providers have pushed to see the 2009 guidelines reduced even further. However, as noted, too-small gains during pregnancy have unacceptable trade-offs in more too-small babies and premature births. The harms are not worth the potential benefits.

Similarly, some care providers have begun to promote the idea of putting non-diabetic obese women on metformin prophylactically to try to prevent big babies. Although this can modestly reduce birthweight in women with gestational diabetes or full-blown diabetes, this study shows that metformin does not reduce birthweight in non-diabetic obese women.

This research effectively disproves the Pederson Hypothesis, which is that big babies are the result of maternal high blood sugar and responding high insulin levels in the baby. Although this feedback loop can cause fetal overgrowth in diabetic women, this study shows that higher birthweights in non-diabetic obese women are NOT because of borderline high blood sugar. There must be something else going on here to cause the bigger babies, which is something I've been saying for years. The authors suspect high lipid levels, but I'm dubious about that one too. Whatever the mechanism is, it's important to go cautiously and avoid jumping to conclusions based only on assumptions about fat people.

This study shows that metformin does not improve outcomes in non-diabetic obese women and should NOT be used for that purpose. Too bad the publicity push around the trial has already created a market for this, and some providers are pushing its use for all their obese patients. Two more trials like this one are already occurring.

Sadly, even though this research was completed and published this past summer, the trial's publicity website has not been updated to show the negative study results or conclusion. Although a bit of publicity about the negative findings has appeared in some medical publications, a large publicity push does not appear to have been done. How many care providers all over the world still have the impression that metformin should be the standard of care of obese women, regardless of blood sugar status?

Researchers, stop promoting a particular approach before the research is even done. A management protocol needs to be proven to be effective and safe in multiple trials before it should be publicized and promoted. Duh.

At least now we know that metformin, while helpful under certain circumstances, is not a cure-all for preventing complications in all high-BMI women or for preventing big babies. Care providers need to restrict its use to situations where it's actually appropriate and needed.



Reference

Lancet Diabetes Endocrinol. 2015 Oct;3(10):778-86. doi: 10.1016/S2213-8587(15)00219-3. Epub 2015 Jul 9. Effect of metformin on maternal and fetal outcomes in obese pregnant women (EMPOWaR): a randomised, double-blind, placebo-controlled trial. Chiswick C1, Reynolds RM2, Denison F1, Drake AJ2, Forbes S2, Newby DE3, Walker BR2, Quenby S4, Wray S5, Weeks A5, Lashen H6, Rodriguez A7,Murray G7, Whyte S1, Norman JE8. PMID: 26165398  Full text available here.

BACKGROUND: Maternal obesity is associated with increased birthweight, and obesity and premature mortality in adult offspring. The mechanism by which maternal obesity leads to these outcomes is not well understood, but maternal hyperglycaemia and insulin resistance are both implicated. We aimed to establish whether the insulin sensitising drug metformin improves maternal and fetal outcomes in obese pregnant women without diabetes. METHODS: We did this randomised, double-blind, placebo-controlled trial in antenatal clinics at 15 National Health Service hospitals in the UK. Pregnant women (aged ≥16 years) between 12 and 16 weeks' gestation who had a BMI of 30 kg/m(2) or more and normal glucose tolerance were randomly assigned (1:1), via a web-based computer-generated block randomisation procedure (block size of two to four), to receive oral metformin 500 mg (increasing to a maximum of 2500 mg) or matched placebo daily from between 12 and 16 weeks' gestation until delivery of the baby...FINDINGS: Between Feb 3, 2011, and Jan 16, 2014, inclusive, we randomly assigned 449 women to either placebo (n=223) or metformin (n=226), of whom 434 (97%) were included in the final modified intention-to-treat analysis. Mean birthweight at delivery was 3463 g (SD 660) in the placebo group and 3462 g (548) in the metformin group. The estimated effect size of metformin on the primary outcome was non-significant (adjusted mean difference -0·029, 95% CI -0·217 to 0·158; p=0·7597). The difference in the number of women reporting the combined adverse outcome of miscarriage, termination of pregnancy, stillbirth, or neonatal death in the metformin group (n=7) versus the placebo group (n=2) was not significant (odds ratio 3·60, 95% CI 0·74-17·50; p=0·11).  INTERPRETATION: Metformin has no significant effect on birthweight percentile in obese pregnant women. Further follow-up of babies born to mothers in the EMPOWaR study will identify longer-term outcomes of metformin in this population; in the meantime, metformin should not be used to improve pregnancy outcomes in obese women without diabetes. 

The Dangers of Estimating Fetal Weight Near Term

Here's a new, interesting study that shows that the routine practice of ordering a fetal weight estimate near term can have negative consequences.

Background: Many care providers order an ultrasound to estimate the baby's weight when the mother is near term, even though research shows this is not a very accurate way to predict the baby's weight.  

Women with babies predicted to be big often are pressured to schedule a planned cesarean to avoid the possibility of shoulder dystocia (shoulders getting stuck, causing injury to the baby), despite little research to show that a cesarean improves outcome.

Or they are encouraged to induce labor early "while the baby is smaller," even though research shows that this actually may increase the risk for cesarean.

Here is a new study that shows that even just doing the fetal-weight ultrasound in the first place may increase the risk for cesarean.

Why?  Probably because knowledge of the results (and nervousness over fetal size) changes the behavior of the care provider so that they have a lower surgical threshold during labor.  

This is particularly an issue for women of size. Even though the majority of "obese" women do not have big babies at all, we do tend to have bigger babies on average than "normal" weight women.  This makes care providers nervous. And since many believe that fat women eat mostly junk food or are practically diabetic anyhow, this increases their fear of big babies even more. So many providers routinely order ultrasounds to estimate fetal weight near term in the pregnancies of obese women.

In this most recent study, 50% of all pregnant women had an ultrasound that calculated Estimated Fetal Weight (EFW) within a month of delivery. Those that did had a higher cesarean rate than those who didn't, especially those whose babies were predicted to be big by EFW.

The rate of EFW ulstraounds is probably even higher in obese women.  How does that then affect our cesarean rates? It is likely another example of management variables that result in increased cesarean rates in obese women.  Yet this is a management variable that can be modified.

An ultrasound to estimate fetal weight near term does not improve outcomes and is a risk factor for an increased cesarean rate.  

References

Am J Obstet Gynecol. 2012 Jul 9. [Epub ahead of print] Estimated fetal weight by ultrasound: a modifiable risk factor for cesarean delivery? Little SE, Edlow AG, Thomas AM, Smith NA.  PMID: 22902073

OBJECTIVE: The purpose of this study was to investigate whether knowledge of ultrasound-obtained estimated fetal weight (US-EFW) is a risk factor for cesarean delivery (CD).
STUDY DESIGN: Retrospective cohort from a single center in 2009-2010 of singleton, term live births. CD rates were compared for women with and without US-EFW within 1 month of delivery and adjusted for potential confounders.
RESULTS: Of the 2329 women in our cohort, 50.2% had US-EFW within 1 month of delivery. CD was significantly more common for women with US-EFW (15.7% vs 10.2%; P < .001); after we controlled for confounders, US-EFW remained an independent risk factor for CD (odds ratio, 1.44; 95% confidence interval, 1.1-1.9). The risk increased when US-EFW was >3500 g (odds ratio, 1.8; 95% confidence interval, 1.3-2.7).
CONCLUSION: Knowledge of US-EFW, above and beyond the impact of fetal size itself, increases the risk of CD. Acquisition of US-EFW near term appears to be an independent and potentially modifiable risk factor for CD.

Am J Obstet Gynecol. 2009 Mar;200(3):340.e1-3. Overestimation of fetal weight by ultrasound: does it influence the likelihood of cesarean delivery for labor arrest? Blackwell SC, Refuerzo J, Chadha R, Carreno CA.  PMID: 19254597

OBJECTIVE: We sought to determine whether the overestimation of ultrasound-derived estimated fetal weight (EFW) is associated with increased diagnosis of labor arrest. STUDY DESIGN: This is a historical cohort study of nulliparous women with term pregnancies who underwent bedside ultrasound examination for EFW before labor induction. Labor outcomes of women with EFW overestimation > 15% the actual birthweight were compared with those with EFW not overestimated. RESULTS: Overestimation of EFW occurred in 9.5% of cases (23/241). The rate of cesarean delivery (CD) for labor arrest was higher for those with EFW overestimation (34.8% vs 13.3%; P = .01) even though there were no differences in length of the induction duration. After adjusting for confounding factors, EFW overestimation remained associated with CD for labor arrest (odds ratio, 4.8; 95% confidence interval, 1.5-15.2). CONCLUSION: Our finding suggests that an overestimation of EFW may be associated with a lower threshold for CD for labor arrest.

Am J Obstet Gynecol. 1995 Oct;173(4):1215-9. Fetal macrosomia: does antenatal prediction affect delivery route and birth outcome? Weeks JW, Pitman T, Spinnato JA 2nd.  PMID: 7485323

OBJECTIVE: Our purpose was to determine whether clinical or ultrasonographic prediction of fetal macrosomia influences subsequent delivery route and birth outcome in a clinical setting where macrosomia is not considered an indication for cesarean delivery. STUDY DESIGN: The hospital records of 504 patients delivered of infants weighing > or = 4200 gm between October 1989 and March 1994 were reviewed. Statistical comparisons were made between patients in whom fetal macrosomia was predicted before delivery (n = 102) and those in whom it was not (n = 402). Cesarean delivery, shoulder dystocia, and birth trauma rates were the variables of interest. RESULTS: Cesarean sections were performed in 52% of the "predicted" group deliveries and in 30% of the "not predicted" group (p < 0.01). The increased cesarean delivery rate in the predicted group appeared to be related to an increased incidence of labor inductions (42.5% vs 26.6%, p = 0.005) and a greater proportion of failed inductions. The proportion of patients delivered by cesarean section without a trial of labor was similar in the predicted and not predicted groups (14.7% vs 10.2%, p = 0.21). There was no significant differences in the incidence of shoulder dystocia or the occurrence of birth trauma. CONCLUSIONS: The antenatal prediction of fetal macrosomia is associated with a marked increase in cesarean deliveries without a significant reduction in the incidence of shoulder dystocia or fetal injury. Ultrasonography and labor induction for patients at risk for fetal macrosomia should be discouraged.

Eur J Obstet Gynecol Reprod Biol. 2002 Oct 10;105(1):20-4. Clinical and ultrasonographic weight estimation in large for gestational age fetus. Weiner Z, et al. PMID: 12270559

OBJECTIVE: To examine prospectively the effect on pregnancy outcome of a management protocol, that adds ultrasonographic weight estimation in fetuses suspected clinically as large. STUDY DESIGN: Prospective follow up study of all singleton deliveries during a 1 year period. All patients underwent routine clinical estimation of fetal weight. When clinical estimation of fetal weight was > or = 3700 g, patients were referred for ultrasonographic estimation of fetal weight. When the latter was > or = 4000 g the patient was informed about the risks of birth trauma. Cesarean section was recommended only when > or = 4500 g. Ultrasonography was repeated every 4 days when possible. Predictive values of clinical and ultrasonographic estimations of fetal weight for diagnosing macrosomia, defined for the purpose of this study as 4000 g or more, and their effect on the rate of cesarean sections. RESULTS: Five hundred fifty-five (14.4%) out of 3844 singletons were estimated as 3700 g or more. Only 315 fetuses had ultrasonographic estimation of weight within 3 days of delivery. The sensitivity of clinical and ultrasonographic prediction of macrosomia was 68 and 58%, respectively. Cesarean section rate in newborns weighing 4000 g or more was 22% when macrosomia was clinically suspected compared to 11% when it was not (P<0.05). In fetuses estimated ultrasonographically as 4000 g or larger the cesarean section rate was doubled (50.7% versus 24.9%, P<0.05) compared to those estimated as smaller than 4000 g, although actual weight of 4500 g or more was recorded in 10.6 and 8.5% of these groups, respectively. There were no cases of shoulder dystocia in macrosomic babies when macrosomia was not detected by ultrasound compared to two cases of shoulder dystocia (2.7%) when macrosomia was detected by ultrasound. CONCLUSION: Antenatal suspicion of macrosomia increased the cesarean section rate while the associated improvement in pregnancy outcome remains questionable. The contribution of ultrasound, added to routine clinical estimation of fetal weight, was clinically insignificant apart from a further increase in cesarean section rate.

Arch Gynecol Obstet. 2008 Sep;278(3):225-30. Epub 2008 Feb 26. Suspected macrosomia? Better not tell. Sadeh-Mestechkin D, et al.  PMID: 18299867

OBJECTIVE: To evaluate the management policy of delivery in a suspected macrosomic fetus and to describe the outcome of this policy. STUDY DESIGN: For this prospective observational study we followed the management by reviewing the medical records of 145 women and their infants. The study population included women at term admitted to the obstetrics department with suspected macrosomic infants, as was diagnosed by an obstetrician and/or by fetal sonographic weight estimation of > or =4,000 g. The comparison group (n = 5,943) consisted of all women who gave birth during the data collection period. RESULTS: Induction of labor and cesarean delivery rates in the macrosomic pregnancies (actual birth weight >4,000 g) of the study group were significantly higher when compared with the macrosomic pregnancies of the comparison group. When comparing the non-macrosomic to the macrosomic pregnancies (actual birth weight </>4,000 g) of the study group no significant difference was demonstrated regarding maternal or infant complications. The sensitivity, specificity and positive predictive value of the methods used for detecting macrosomia were 21.6, 98.6 and 43.5%, respectively. CONCLUSION: Our ability to predict macrosomia is poor. Our management policy of suspected macrosomic pregnancies raises induction of labor and cesarean delivery rates without improving maternal or fetal outcome.

Am J Obstet Gynecol. 2006 Sep;195(3):657-72. Epub 2006 Apr 21. Shoulder dystocia: the unpreventable obstetric emergency with empiric management guidelines. Gherman RB, et al.   PMID: 16949396

OBJECTIVE: Much of our understanding and knowledge of shoulder dystocia has been blurred by inconsistent and scientific studies that are of limited scientific quality. In an evidence-based format, we sought to answer the following questions: (1) Is shoulder dystocia predictable? (2) Can shoulder dystocia be prevented? (3) When shoulder dystocia does occur, what maneuvers should be performed? and (4) What are the sequelae of shoulder dystocia? STUDY DESIGN: Electronic databases, including PUBMED and the Cochrane Database, were searched using the key word "shoulder dystocia." We also performed a manual review of articles included in the bibliographies of these selected articles to further define articles for review. Only those articles published in the English language were eligible for inclusion. RESULTS: There is a significantly increased risk of shoulder dystocia as birth weight linearly increases. From a prospective point of view, however, prepregnancy and antepartum risk factors have exceedingly poor predictive value for the prediction of shoulder dystocia. Late pregnancy ultrasound likewise displays low sensitivity, decreasing accuracy with increasing birth weight, and an overall tendency to overestimate the birth weight. Induction of labor for suspected fetal macrosomia has not been shown to alter the incidence of shoulder dystocia among nondiabetic patients. The concept of prophylactic cesarean delivery as a means to prevent shoulder dystocia and therefore avoid brachial plexus injury has not been supported by either clinical or theoretic data....