Pam. Weiss et al., Special aspects of insulin therapy in pregnancies complicated by gestational diabetes mellitus (GDM), GEBURTSH FR, 60(7), 2000, pp. 366-379
Objective: In pregnancies complicated by gestational diabetes mellitus (GDM
) about 10-20% of fetuses develop hyperinsulinism. Aggressive insulin thera
py is required to avoid early and late complications for fetal hyperinsulin
ism in these offspring. There is no universally agreed upon indication for
insulin therapy. The indication for insulin can be based on maternal parame
ters, such as fasting blood glucose, postprandial blood glucose or mean blo
od glucose (MBG) values. It can also be initiated prophylactically or be ba
sed on fetal parameters, such as macrosomia or elevated amniotic fluid insu
lin.
Methods: Suggested maternal glucose limits For the indication of insulin th
erapy are 95 - 105 mg/dl (5.3 - 5.8 mmol/l) for fasting blood glucose, 120
- 130 mg/dl (6.7 - 7.2 mmol/l) for postprandial blood glucose, and 90 - 108
mg/dl (5 - 6 mmol/l) for MBG. Prophylactic insulin therapy is administered
independently of maternal glycemia. Macrosomia has been defined as the 75t
h percentile of fetal abdominal circumference; the upper limit of normal am
niotic fluid insulin is 8 mu U/ml (48 pmol/l). Maternal glycemia correlates
poorly with fetal hyperinsulinism because of individual variations in plac
ental transport and the placentas barrier function. Accordingly, the matern
o-fetal glucose gradient can vary widely. Additionally, the fetal islet org
ans vary widely in sensitivity to glucose stimuli.
Results: After the 31st gestational week hyperinsulinemic fetuses syphon of
f maternal glucose and thus reduce maternal postprandial glucose levels by
20 mg/dl (1.1 mmol/l) on average. Consequently, it cannot be distinguished
whether maternal euglycemia is a result of normal glucose tolerance, adequa
te treatment, or the fetoplacental glucose steal phenomenon. A policy of pr
ophylactic insulin treatment of all pregnancies complicated by GDM would ov
ertreat the 80 - 90% of fetuses who are normoinsulinemic. A policy based on
sonographic criteria of macrosomia would treat only the hyperinsulinemic f
etuses who are also macrosomic, as well as large but normoinsulinemic fetus
es. Also, hyperinsulinism precedes macrosomia by weeks or months so that in
sulin treatment is late. Insulin treatment based on measurement of insulin
levels in the amniotic fluid avoids undertreatment and overtreatment. Compl
ications of amniocentesis are negligible but patient acceptance is only 80
- 90%. We analyzed the extent of undertreatment and overtreatment with the
fetus as a sensor in 542 GDM with known amniotic fluid insulin levels.
Conclusions: Assuming that insulin therapy is only necessary in GDM with fe
tal hyperinsulinism, no undertreatment ensues from prophylactic insulin the
rapy or from insulin treatment based on amniotic fluid insulin levels. At a
MBG of 90 - 108 mg/dl or estimating the 75th weight percentile for indicat
ion, undertreatment results in 16 - 58% or 44%, respectively. Insulin treat
ment based on a MBG of 90 - 108 mg/dl, the 75th weight percentile and proph
ylactic insulin therapy lead to overtreatment in 75 - 18%, > 25% and 100% o
f normoinsulinemic cases, respectively. The problem is undertreatment becau
se preventable sequelae, such as premature births, cesarean deliveries and
diabetes in the adolescent ensue from untreated fetal hyperinsulinism. Cons
equently a policy of insulin treatment for patients with a MBG greater than
or equal to 108 mg/dl (greater than or equal to 6 mmol/l) leads to costs e
ight times higher than those of overtreatment with prophylactic insulin. Du
e to insulin resistance, insulin requirements in GDM are high. Moreover, in
sulin administration reduces the mother's own insulin secretion by 30 U/24
h. Consequently the insulin requirement needed to overcome insulin resistan
ce, the reduction of maternal insulin synthesis, and fetal hyperinsulinism
is 0.8 - 1.2 U/24 h. Because maternal postprandial insulin secretion is del
ayed in GDM, insulin administration at the main meals is essential to preve
nt postprandial hyperglycemia. Consequently, a basal-bolus schema is approp
riate. Administration of human insulin is preferable in order to prevent de
velopment of insulin antibodies.