I've downloaded the trial and reviewed it:
Women with normal fasting blood sugar and elevated blood sugar after eating (N=958) were randomly allocated to treatment or usual care. As one would expect, women were mostly multiparous (75-79%) and had high BMI (mean 30). To partially mask the control group, investigators also included 931 women with normal glucose tolerance matched for race and BMI above or below 27 in the control group. Their outcomes were not reported, so I am not sure what effect their inclusion had on the care of control group women.
Treated women were managed according to American Diabetes Assoc. guidelines and were given insulin if blood sugars were not normalized. Few women required insulin (8%). The short version of the ADA recommendations (I obtained the guidelines the investigators referenced) are essentially a healthy diet and exercise, and while high BMI women should not be put on a weight-loss diet, calories can be restricted somewhat to limit weight gain with monitoring to ensure avoidance of low blood sugar. No information is given on details of usual care.
No babies were stillborn or died neonatally, and birth trauma rates were similar, but a much larger study would be needed to detect differences between groups. Babies in the treatment group wre less likely to weigh over 4000 g (8 lb 13 oz) (6% vs. 14%) and to be in the upper 10% of weight for their gestational age (7% vs. 15%) without increasing likelihood of being small for gestational age--an important point because aggressive sugar control can result in more undersized babies.
Treated women gained less weight after trial enrollment (mean 3 kg vs. 5 kg or 6.5 lb vs. 11 lb), had a lower BMI at delivery (31 vs. 32), and their babies were less likely to experience shoulder dystocia (head is born but shoulders hang up behind the pubic bone) (1.5% vs. 4%). Treated women were less likely to have c/secs both overall (27% vs. 34%) and after excluding women with conditions unrelated to gestational diabetes that would increase likelihood of c/sec such as prior c/secs or abnormal presentation (13% vs. 20%). They were also less likely to experience pregnancy hypertension (9% vs. 14%) and preeclampsia (2.5% vs. 5.5%).
What do we learn from this? If you take a group of high-BMI women, give them a strong motivation to eat right (probably under the guidance of a registered dietician since these were academic institutions) and exercise more, and monitor them closely, you will see the benefits of that compared with when you don't. And if their obstetricians both believe that treatment works and see differences that support the belief that the baby will be deliverable vaginally, such as less weight gain, they will be less likely to operate. (Lord knows we have enough studies showing that clinician belief is the primary factor affecting c/sec rates.) So what's new? I have always maintained that if all GD screening did was identify women with what should better be called "carbohydrate intolerance" and helped them with diet and lifestyle, it would be a good thing, and I would be all for it. Kudos to these investigators for confirming that. I maintain, though, first, that they probably could have accomplished the same thing merely by treating women according to BMI, which would also have benefited normoglycemic high BMI women, and second, that a trial carried out in a group of academic institutions does not reflect the way the typical community ob identifies and treats GD--and probably not even academic institutions not participating in a trial of GD treatment--to the detriment of women so labeled.
-- Henci