Aust. N Z J Obstet Gynaecol 1996: 36: 3: 248

Impaired Glucose Tolerance in Pregnmcy - Is It of Consequence?

Y.Y. Tan', and G.S.H. Yeo' Department of Maternal and Fetal Medicine, Kundung Kerbau Hospifal, Singapore

Summary: This study was done to determine if impaired glucose tolerance in pregnancy was associated with increased maternal and neonatal morbidity and if so, whether the increased morbidity was due to the confounding factors of increased maternal age and maternal obesity. It was a retrospective analysis to compare 944 women with impaired glucose tolerance (IGT) in pregnancy with 10,065 women without abnormal glucose tolerance. The incidence of impaired glucose tolerance in pregnancy was 8.6% in this study. Even when maternal age and obesity were excluded, the IGT group had significantly higher risks of labour induction (relative risk, RR, 1.15); Caesarean section (RR: overall 1.43, elective 1.72, emergency 1.31); Caesarean section for dystocidno progress (RR 1.60); macrosomia (RR 1.69,1.76, 1.61 for birth-weight >=97th, 95th, 90th percentiles respectively) and shoulder dystocia (RR 2.84) when compared to the nondiabetics (NDM). The risks of hypertensive disease (RR 1.22) and Caesarean section for fetal distresshhick meconium-stained liquor (RR 1.53) were also higher in the IGTgroup but these increases were not statistically significant when maternal age and obesity were excluded. There was no significant difference in the rates of low A p g ~ scores at 1 and 5 minutes between the 2 groups.

Is impaired glucose tolerance in pregnancy of any consequence? This is a highly controversial subject. Tallarigo et a1 (1) studied women with normal results on testing with a 100 g-load oral glucose tolerance test and found that those with higher 2-hour plasma glucose levels had significantly higher rates of macrosomia, congenital abnormalities, toxaemia and Caesarean section. This indicates that mild degrees of hyperglycaemia, not just frank diabetes, in pregnancy are associated with increased risks to both mother and fetus.

Abell (2) stated that 'the significance of less overt forms of glucose tolerance abnormality in pregnancy has been clearly elucidated, but it is disturbing to sense that there continues to be less than enthusiastic reception of this fact' and that 'gestational diabetes demands to be diagnosed during pregnancy, leading to careful pregnancy and delivery management, with continued follow-up of the patient for life'.

On the other hand, Ales and Santini (3) maintained that 'although gestational glucose intolerance is

1 . Registrar. 2. Senior Consultant. Address for correspondence: J h Yin-Yin Tan. Department of Maternal Fetal Medicine, Kandang Kerbau Hospital, 2 Hampshire Road, Singapore 219428.

associated with the remote development of diabetes meilitus, the risk to the mother during the index pregnancy and the risk to her fetus remain uncertain' and that 'the scientific data supporting a universal screening programme - showing treatment of gestational glucose intolerance does more good than harm - are limited'. Jarrett (4) reviewed the subject and concluded that 'any maternofetal morbidity is more Iikely to be due to maternal age or obesity or to the effects and consequences of diagnosis than t o the glucose intolerance' and that 'screening for gest- ational impaired glucose tolerance does not accord with any of the usual requirements for it worthwhile screening programme'.

Opinions vary widely, from one extreme t o the other. Hence, we did this study to determine if impaired glucose tolerance during pregnancy was associated with higher risks to the mother and the neonate. If so, was the increased maternal and neo- natal morbidity due to the confounding factors of maternal age and maternal obesity or due to the state of glucose intolerance per se?

SUBJECTS AND METHODS This is a retrospective analysis of pregnancies

complicated by impaired glucose tolerance (IGT). Impaired glucose tolerance was detected by the screening of a high-risk obstetric population with any 1 of the following risk factors:

Y.Y. TAN ANV G.S.H. YEO 249

1. Family history of diabetes mellitus in first degree

2. Past obstetric history of: relatives.

(a) gestational diabetes mellitus (b) a big baby with a birth-weight of 4 kg or more (c) recurrent miscarriages or unexplained perinatal

death.

(a) glycosuria (b) abnormal glucose challenge test result (c) a big fetus (d) polyhydramnios (e) recurrent urinary or genital tract infection.

(a) maternal age of 35 years or more (b) maternal obesity - weight >90th percentile (c) use of diabetogenic drugs such as betasympath-

The diagnosis of impaired glucose tolerance was based on a 75g-load oral glucose tolerance test and criteria defined by the World Health Organization in 1985, that is, 2-hour postload blood sugar level of 7.8 mmoUL (140 mg/dL) or more but less than 11.1 mmoVL (200 mg/dL).

The women with impaired glucose tolerance were put through blood sugar profile monitoring. Seven- point blood sugar profiles, consisting of 3 sets of pre- prandial and 2-hour postprandial and 10 pm glucose levels, were done for women who were admitted. For those who were assessed at our Obstetric Day Assess- ment Centre (ODAC), 6-point blood sugar profiles (excluding the 10 pm glucose) were done. Those with unsatisfactory blood sugar profiles were then admitted for control.

These women were put on dietary control with a 30- 35 kcalkg ideal body weight diet either concurrently with or after the blood sugar profile. The patients were taught self glucose monitoring at the same time. Insulin therapy was instituted when the preprandial glucose level exceeded 5.5 mmoUL or when the 2-hour postprandial glucose level exceeded 7.0 mmoyL.

Patients were discharged after satisfactory gly- caemic control was achieved. They were then instructed to check their blood sugar profiles at home. Patients were reviewed regularly, either by their own obstetrician or at the Diabetic Clinic, which was jointly run by an endocrinologist and an obstetrician. Treatment was adjusted accordingly.

Results of all oral glucose tolerance tests done in the hospital were recorded in a computer database. Data on antenatal complications were collected when patients were admitted to the labour ward. After delivery, ail data was entered into the labour ward computer database. The oral glucose tolerance test results were then matched with the labour ward database.

3. Pregnancy complications such as:

4. Constitutional factors such as:

omimetics and corticosteroids.

The maternal and neonatal outcomes were com- pared between women with impaired glucose tolerance (IGT) and the nondiabetics (NDM). the nondiabetics included:

1. Those who were not established diabetics and who did not have any glucose tolerance test, generally because of the absence of risk factors

2. Those who were selected for oral glucose toler- ance testing but had normal results.

Babies weighing less than 500 g were excluded from the comparison of neonatal outcome and still- births were excluded from the comparison of Apgar scores.

Subgroup analyses were then done, excluding, in turn: 1 . Women aged 35 years or more 2. Obese women whose weight at the last antenatal

visit were at or above the 90th percentile for this population

3. Women who were either obese or aged 35 years or more. The relative risks (RR) and the 95% confidence

intervals (CI) in the various subgroups were cal- culated and compared graphically.

RESULTS There was a total of 1 1,038 deliveries from April 5,

1994 to December 3 1, 1994. Of these, there were 944 women with impaired glucose tolerance. Hence, the incidence of impaired glucose tolerance (IGT) in pregnancy was 8.6%.

Overall comparison between women with IGT and ihe non-diubetics (NDM)

Our definition of hypertensive disease included preeclampsia, pregnancy-induced hypertension and chronic hypertension. The relative risk (RR) of hypertensive disease in the IGT group was 2.43 that for the nondiabetics (NDM) (table 1).

Women with IGT were 1.18 times more likely to have induction of labour and 153 times more likely to have Caesarean section. The relative risks of elective and emergency Caesarean section in the IGT group were 1.98 and 1.32 respectively (table 1). On examining the indications for Caesarean section, the IGT group had significantly higher rates of Caesarean section for disproportion or no progress in labour (RR 1.48); fetal distress or thick meconium-stained liquor (TMSL) (RR 1.64); and previous Caesarean section (RR 1.88). The duration of labour in women requiring emergency Caesarean section were compared in the 2 groups. The mean duration of labour in the IGT group was 11.92h (SE 1.33 hours) while that in the NDM group was 11.27h (SE 0.27 hours).

Our definition of macrosomia was based on the study population. The 97th, 95th and 90th percentiles of birth-weights of babies delivered at or after 37

250 AUST. AND N.Z. JOURpjAL OF OBSTETRlCS AND GYNAECOLoGY

weeks’ gestation were determined for a11 deliveries during the study period. The IGT group had significantly higher rates of macrosomia, whichever definition was used. The relative risks were 2.16, 1.90 and 1.69 respectively when the 97th, 95th and 90th percentiles were used as cut-off points (table 1). Associated with this was a significantly higher rate of shoulder dystocia (RR 3.28).

Table 1. Comparison of Complications in the IGT and NDM Groups Overall

IGT NDM RR 95%CI n (%) n 1%)

Maternal morbidity (n=944) (n=10,065)

Hypertension 63 (6.7) Labour induction 432 (45.8) Caesarean section - Total 233 (24.7) - Elective 94 (10.0) -Emergency 139 (14.7) CS indication -DistresslTUSL 36 (3.8) - Dystocia 69 (7.3) -Previous CS 68 (7.2) Neonatal morbidity

Major birth defect 3 (0.3) Perinatal mortality 0 (0.0)

Macrosomia (n=944) - BW 297% 52 (5.5) - BW 295% 80 (8.5) - B W m 145 (15.4) Shoulder dystocia 12 (1.3)

(n=944)

276 (2.7) 3,892 (38.7)

1,627 (16.2) 507 (5.0)

1,120 (11.1)

234 (2.3) 497 (4.9) 385 (3.8)

(n=10.065) 35 (0.3) 21 (0.2) (n= 10,060)

257 (2.6) 449 (4.5) 914 (9.1) 39 (0.4)

2.43 1.18

1.53 1.98 1.32

1.64 1.48 1.88

0.91

2.16 1.90 1.69 3.28

1.87-3.17 1.10-1.27

1.36-1.72 1.60-2.44 1.12-1.56

1.16-2.32

1.47-2.42 1.16-1.89

0.28-2.97

1.61-2.88 I .51-2.39 1.44- 1 .w 1.72-6.24

Apgar 4 at (n=939) (n=9,970)

Apgar <5 at 5 minutes 1 (0.1) 17 (0.2) 0.62 0.08-4.69 IGT = impaired glucose tolerance: NDM = nondiabetic; RR = relative risk: CI = confidence interval: CS = Caesarean section: TMSL = thick meconium-stained liquor; BW = birth-weight

1 minute 16 (1.7) 152 (1.5) 1.12 0.67-1.86

On the other hand, the risk of poor Apgar scores (4) at 1 and 5 minutes was not significantly different in the 2 groups (table 1).

The incidence of major birth defects and perinatal mortality were reported for completeness sake. However, the numbers for these 2 outcomes were too small for any meaningful analyses (tables 1-4).

Subgroup analysis excluding women aged 35 years or more

Maternal and neonatal morbidity were still sig- nificantly increased in the IGT group when compared to the NDM group even when older women were excluded from analysis (table 2) . The only exception was the risk of Caesarean section for fetal distress/ thick meconium-stained liquor which, though higher in the IGT group (RR 1.54), did not reach statistical significance.

Table 2. Comparison of Complications in the ICT and NDM G m p s After Excluding Women Aged 35 Years or More

IGT NDM RR 95% CI n (%) n (%I

Maternal morbidity (n=620) (n=8,573)

Hypertension 29 (4.7) 204 (2.4) 1.97 1.34-2.88 Labour induction 296 (47.7) 3,416 (39.8) 1.20 1.10-1.31 Caesarean section - Total 133 (21.5) 1,283 (15.0) 1.43 1.22-1.68 - Elective 47 (7.6) 379 (4.4) 1.71 1.28-2.30 -Emergency 86 (13.9) 904 (10.5) 1.32 1.07-1.62 CS indication

- Dystocia 49 (7.9) 410 (4.8) 1.65 1.24-120 -Previous CS 37 (6.0) 278 (3.2) 1.84 1.32-2.57

-Di~tresslTUSL 21 (34) 188 (2.2) 1.54 0.99-2.41

Neonatal morbidity

Major birth defect 0 (0.0) Perinatal mortality 0 (0.0)

Macrosomia - BW 297% 35 (5.61 - BW 295% 57 (9.2)

Shoulder dystocia 10 (1.6)

Apgar 4 at 1 minute 9 (1.5)

(n=620)

(n=620)

-BWSO% 104 (16.8)

(n-617)

(n=8,573) 29 (0.3) 15 (0.2) (114,568)

207 (2.4) 2.34 1.65-3.31 363 (4.2) 2.17 1.66-2.83 746 (8.7) 1.93 1.60-2.33 32 (0.4) 4.32 2.13-8.74 (n=8,501)

122 (1.4) 1.02 0.52-1.99 Apgar <5 at 5 minutes 1 (0.2) 15 (0.2) 0.92 0.12-6.94 IGT = impaired glucose tolerance: NDM = nondiabetic; RR = relative risk: CI = confidence interval: CS =Caesarean section: TMSL= thick meconium-stained liquor: BW = birth-weight

Table 3. Comparison of Complications in the IGT and NDM Groups After Excluding Women with Weights at 90th

Percentile or More IGT NDM RR 95QCI

n (F) n (%)

Maternal morbidity (n=758) (n=8.448)

Hypertension 34 (4.5) 186 (2.2) 2.04 1.42-2.92 Labour induction 336 (44.3) 3.349 (39.6) I . I2 1.03- I .22 Caesarean section -Total 184 (24.3) 1.321 (15.6) 1.55 1.36-1.78 -Elective 75 (9.9) 419 (5.0) 1.99 1.58-2.52 -Emergency 109 (14.4) 902 (10.7) 1.35 1.12-1.62 CS indication -DistresslTMSL 27 (3.6) 190 (2.2) 1.58 1.07-2.35 - Dystocia 53 (7.0) 409 (4.8) 1.44 1.10-1.90 -Previous CS 56 (7.4) 304 (3.6) 2.05 1.56-2.70 Neonatal morbidity

(n=758) (~8,448) Major birth defect 2 (0.3) 32 (0.4) 0.68 0.16-2.85 Perinatal mortality 0 (0.0) 10 (0.1)

(n=758) (nd3.446) Macrosomia - BW 297% 27 (3.6) 170 (2.0) 1.77 1.19-2.64 - BW 295% 48 (6.3) 317 (3.8) 1.69 1.26-2.27 - BW 290% 93 (12.3) 684 (8.1) 1.52 1.24-1.86 Shoulderdystocia 6 (0.8) 30 (0.4) 2.23 0.93-5.34

(n=754) (n=8,398) Apgar 4 at 1 minute 13 (1.7) 110 (1.3) 1.32 0.74-2.33 Apgar <5 at 5 minutes I (0.1) 13 (0.2) 0.86 0.11-6.54 IGT = impaired glucose tolerance: NDM = nondiabetic: RR = relative risk CI = confidence interval: CS = Caesarean section; TMSL = thick meconium-stained liquor: BW = hinh-weight

Y.Y. TAN AND G.S.H. Y w 25 1

- , __ Hypertensive 0 Disease A

W --- - I ' -- I M W

0 1 - 1 Labour inducuon A

W

M W

t -

0 t - r - i Overall

Secnon Caesarean A 1 1 , -1

W

M W

0 Electtwe Camrarean A Sacnon

i z; ~ i i

- T ----- i W i / T i

M W ' 1 , - 4

Emargancy 0 t Caesarean A I SecUon 4

W 1 , , i

M W i 0 1 2 3

0 = overall, A = excluding women aged 35 years or more; W = excluding women WK weights a t 90th percentile or more, A and W = excluding women aged 35 years or more wrth weights a t 90th percentile or more

Figure 1. Companson of relative nsks and 95% confidence intervals of vanous maternal complications in vanous groups of women with impaired glucose tolerance.

Fetal Disters

0

/TMSL A t------------i i

W I -~ .

A M

Birth- Welght >=97% A

T . -7

,

I

Birth- ' 1 Weight >=G5% A l

- w ,

B1Rh- Waight >=GO% A

W

M W

0 1 2 3 0 = overall; A = excluding women aged 35 years or more: W = excluding women with welghts at 90th percentile or more, A and W = excluding women aged 35 years or more with weights at 90th percentile or more

Figure 3, Companson of relauve risks and 95% confidence intervals of macmsomia (birth-wetghts above 97th, 95th. 90th percentiles) in vanous groups of women with impaued glucose tolerance.

Shoulder 0 1 Dystocia I

Dystocia / No Progress A

, Aaw

I I

Previous 0 Caesarean Section

A

I !

, I--- 7

W I------ I

____--- AaW - . -

0 1 2 3 0 = overall; A = excluding women aged 35 years or more; W = excluding women with weights a t 90th percentile or more: A and W = excluding women aged 35 years or more with weights at 90th percentile or more

Figure 2. Comparison of relative risks and 95% confidence intervals of various indications for Caesarean section in various groups of women with impaired glucose tolerance.

~

I <5 at Apgar 5 Min O ' ~

I A - t -- - -

w ,- 4- ____- --

+&# .- , - - - .- - - --__

0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 0 = overall. A = excluding women aged 35 years or more; W = excluding women with weights a t 90th percentrle or more: A and W = excluding women aged 35 years or more w K weights a t 90th percentile or more

Figure 4. Companson of relative nsks and 95% confidence intervals of shoulder dystocia and low Apgw scores at 1 and 5 minutes in vanous groups of women with impaired glucose tolerance.

252 AVST. 4 N D N.Z. JOL'RNAL OF OBSTETRICS 4 S D GYNAECOLOGY

Compared to the whole group of women with IGT (group O), the younger women with IGT (group A) tended to have slightly lower risks of hypertensive disease and elective Caesarean section (figure I). On the other hand, these younger women had higher risks of macrosomia (figure 3), shoulder dystocia (figure 4) and Caesarean section for dystocia (figure 2). How- ever, these differences were not statistically significant.

Subgroup analysis excluding women with weights at 90th percentile or more

Comparing the 2 groups of nonobese women, those with IGT still had significantly higher maternal mor- bidity and macrosomia than those without diabetes (table 3). Although the relative risk of shoulder dystocia was 2.23 that in the NDM group, this dif- ference did not reach statistical significance (table 3).

When compared to the whole group of women with IGT (group 0), this group of nonobese women (group W) had slightly lower risks of hypertensive disease, induction of labour (figure l), macrosomia (figure 3) and shoulder dystocia (figure 4).

Subgroup analysis excluding older and obese women Compared to the NDM group, the younger and

nonobese women with IGT had significantly higher maternal and neonatal morbidity except for hyper- tensive disease and Caesarean section for fetal distresskhick meconium-stained liquor (table 4).

Table 4. Comparisons of Complications in the IGT and NDM Groups After Excluding Women Aged 35 Years or More or

With Weights at 90th Percentile or More IGT NDM RR 95%CI

n (9) n (%)

M a t e d morbidity (n=489) (n=7,226)

Hypertension 1 1 (2.2) 133 (1.8) 1.22 0.67-2.25 Labour induction 229 (46.8) 2,938 (40.7) 1.15 I .04- I .27 Caesarean section -Total 101 (20.7) 1.043 (14.4) 1.43 1.19-1.72 - Efective 36 (7.4) 310 (4.3) 1.72 1.23-2.39 -Emergency 65 (13.3) 733 (10.1) 1.31 1.03-1.66 CS indication -Distress/IuSL 16 (33) 155 (2.1) 1.53 0.92-2.53

-PreviousCS 28 (5.7) 215 (3.0) 1.92 1.31-2.82 Neonatal morbidity

(11489) (n=7,226) Major birth defect 2 (0.0) 27 (0.4) Perinatal mortality 0 (0.0) 7 (0.1)

( n 4 8 9 ) (n=7,224) Macrosomia - BW 3 7 6 16 (3.3) 1 4 0 (1.9) 1.69 1.01-2.81 - BW 3 5 9 31 (6.3) 260 (3.6) 1.76 1.23-2.53

Shoulderdystocia 5 (1.0) 26 (0.4) 2.84 1.10-7.37

Apgar <5 at (n=487) 1 minute 8 (1.6) 89 (1.2) 1.33 0.65-2.72 Apgar <5 at 5 minutes 1 (0.2) 10 (0.1) 1.48 0.19-11.52 IGT = impaired glucose tolerance; NDM = nondiabetic; RR =relative risk: CI =confidence interval: CS = Caesarean section: TMSL = thick meconium-stained

- Dystocia 37 (7.6) 341 (4.7) 1.60 1.16-2.22

- B W S " G 62 (12.7) 568 (7.9) 1.61 1.26-2.06

(n=7,188)

Compared to the whole IGT group (group O), the younger and nonobese women with IGT (group A&W) had lower risks of hypertensive disease though the difference was not statistically significant.

DISCUSSION Incidence

In this study, the incidence of impaired glucose tolerance in pregnancy was 8.6%, which can hardly be considered insignificant. This incidence is also much higher than those reported in earlier local studies. Ng et al (5) found that incidence of gestational diabetes was 0.93% in a study from 1978 to 1980, where women with risk factors were given a 50g-load glucose tolerance test. Kek et a1 (6) found a similar incidence of gestational diabetes of 0.99% in a 1981 study using the same screening process and diagnostic test. In a more recent study by Chan et a1 (7) where risk factor screening followed by a 75g-load glucose tolerance test was used, the incidence of impaired glucose tolerance in pregnancy was 1.93%. In contrast, Wang et al (8) screened all women attending an antenatal clinic with a 75g-load glucose tolerance test, regardless of the absence of risk factors, and found that the incidence of impaired glucose tolerance in pregnancy was 10.3% and that 50% of these patients did not have any of the classical risk factors for gestational diabetes. Our higher incidence as compared to the studies in the 70s and 80s may, in part, be due to the higher glucose load used. In addition, during our study period, some primary healthcare centres were screening patients with the 50 g glucose challenge test which was previously not used here. The addition of this extra risk factor, an abnormal glucose challenge test result, in the screen- ing process may have accounted for the incidence being higher when compared to Chan's (7) study. However, our incidence is consistent with Wang's ( 8 ) study as we may still be missing out a proportion of patients without risk factors as universal screening was not practised.

Hypertensive disease Is hypertensive disease more prevalent in women

with impaired glucose tolerance? Al-Shawaf et a1 (9) found that the incidence of hypertension in the IGT group was 14 of 177 (7.9%), which was higher than in the control group (RR 1.58). However, this did not reach statistical significance. Roberts et al ( 10) found a significantly higher rate of pregnancy-induced hypertension of 12.5% in those with IGT compared to 8.4% in those with normal glucose tolerance (NGT) (RR 1.49). We also found an increase in the rate of hypertensive disease in pregnancy in the IGT group (RR 2.43) which was statistically significant.

It has been argued that the higher incidence of hypertensive disease in women with IGTmay be

Y.Y. TAN AND G.S.H. YEO 253

related to the higher prevalence of increased maternal age and obesity in these women instead of the state of glucose intolerance per se. In our study, hypertensive disease remained significantly increased in the IGT group when older women alone and when obese women alone were excluded in turn. However, when both older and obese women were excluded from analysis, the difference was no longer statistically significant even though the relative risk (1.22) was increased (table 4). This may indicate that maternal age and obesity are important confounding factors where hypertensive disease is concerned. Impaired glucose tolerance per se did not seem to significantly increase the risk of hypertensive disease in pregnancy in younger, nonobese mothers in our study. In contrast, Pettitt et al ( 1 1 ) found that toxaemia of pregnancy was strongly associated with third trimester plasma glucose level, with a steep increase in rate beyond a glucose level of 160 mg/dL. Although they found that the rate of toxaemia varied with weight and age, binary multiple regression analysis showed that the effect of glucose intolerance persisted after accounting for age and weight. In our study, chronic hypertension was not excluded in the analysis of hypertensive disease in pregnancy. This could have masked the effect of glucose intolerance on preeclampsia alone.

Obstetric in fentention The induction and overall Caesarean section rates

were significantly higher in the IGT group in our study, with relative risks of 1.18 and 1.53 respectively. This is similar to Roberts et al’s (10) study where the induction rate in the IGT group (42.5%) was 1.32 times that in the NGT group (32.3%). They also found a significantly higher total Caesarean section rate in the IGT group (24.2%) than in the NGT group (14.4%) (RR 1.68). This was largely due to an increase in the emergency Caesarean section rate. Roberts et a1 (10) did not find any significant difference in elective Caesarean section rates in the 2 groups whereas we had a significantly higher rate in the IGT group. On the other hand, Al-Shawaf et a1 (9) did not find any increase in Caesarean section rate in the IGT group (6.8%), compared to 9% in the unit. Li et a1 (12) put women with IGT and NGT together, blinded the results and randomized them into a treatment and a control group. They did not find any difference in the induction or Caesarean section rates between the treatment and control groups. However, there was no comparison of the rates between the IGT and the NGT groups.

In our study, the risk of obstetric intervention in terms of labour induction and Caesarean section, whether elective or emergency, remained significantly increased in women with impaired

glucose tolerance even after excluding the ol&r and obese women. Pettitt et a1 (11) also found that Caesarean section rates increased with increasing third trimester glucose levels. Roberts et a1 (10) found that the overall Caesarean section rate was signifi- cantly higher in women with IGT compared to those with normal glucose tolerance (NGT). In their study, maternal weight was not a confounding factor as this was comparable in the 2 groups.

Disproportion and fetal distress Roberts et a1 (10) found that their higher emergency

Caesarean section rate in the IGT group was due to a small and nonsignificant increase in the incidence of both cephalopelvic disproportion and fetal distress. In our study, the rate of Caesarean section for fetal distresdthick meconium-stained liquor was not sig- nificantly higher in the IGT group after exclusion of the older and obese women. However, the rate of Caesarean section for dystocidno progress in labour remained significantly increased in the IGT group despite excluding the older and obese women. Know- ledge of the diagnosis of impaired glucose tolerance could have biased the clinician’s management and the labelling of the indication for Caesarean section. However, this did not seem to be an important confounding factor in our study as there was no significant difference in the mean duration of labour in women requiring emergency Caesarean section in the 2 groups - 11.92 hours (SE 1.3 hours) for the IGT group and 11.27 hours (SE 0.27 hours) for the NDM group. We would expect the mean duration of labour to be shorter in the IGT group if the diagnosis had biased the clinician’s management. Hence, the state of impaired glucose tolerance itself is associated with an increased risk of Caesarean section for dystocidno progress in labour. This is hardly surprising as we found a significant increase in the risk of macrosomia in the IGT group.

Macrosomia As the definition of macrosomia has not been

standardized, we examined 3 commonly used criteria: the 97th, 95th and 90th percentiles €or birth-weight of term babies in our study population. Interestingly, we found that our 97th percentile was equivalent to 4,000 g, which has often been used to define macrosomia. The IGT group had higher rates of macrosomia regardless of which definition we used. The reiative risks of macrosomia in women with IGT increased (1.69, 1.90, 2.16) the stricter the definition of macrosomia (Wh, 95th, 97th percentiles).

Al-Shawaf et al (9) found a significantly higher mean birth-weight in the IGT group (3,427 g) compared to the control group (3,320 8). The rate of macrosomic babies (>= 4kg) in the 1GT group

254 AGST. AND N.Z. JOURNAL OF OBSTETRICS AND GYNAECOL~GY

(10.7%) was higher than in the control group (6.7%) but this did not reach statistical significance. Johnstone et a1 (13) also found increased rates of macrosomia (>=4.5 kg} in the IGT group (7.0%) compared to the control group ( 1.8%).

Green et a1 (14) stated that ‘because the prevalence of obesity is increased among women with gestational diabetes, fetal overgrowth may be attributable at least in part to maternal obesity’ in this group of women. Pettitt et al ( 1 1 ) found that the percentage of pregnancies with large-for-gestational age (LGA) infants increased with increasing third trimester glucose levels, with a steep rise beyond a glucose level of 100 mg/dL and an even steeper rise beyond a glucose level of 160 mg/dL. Their LGA rate also varied with maternal weight and age. Within strata of increasing maternal weight or age, LGA rates increased with the glucose level except in the heaviest and oldest categories. However, after maternal age and weight were accounted for in a binary multiple regression analysis, third trimester glucose level was no longer significantly associated with the LGA rate.

In Roberts et al’s (10) study, maternal weight was not a confounding factor as both groups of women were selected based on risk factor screening which included a past history of macrosomia and maternal weight >90 kg. Despite this, the IGT group had higher mean birth-weight (3,620 g in IGT, 3,554 g in NGT) and macrosomia rate, though these did not reach statistical significance. Similarly, Weeks et al (1 5) found that women with gestational diabetes had a higher rate of macrosomia and in a subanalysis in which obese women were excluded, nonobese women with gestational diabetes remained at increased risk of macrosomia. Li et al’s (1 2) randomized controlled trial compared treatment versus no treatment in women with IGT. They found that in those who were not treated, the mean birth-weight (3,407 g) was significantly higher than in those who were treated (3,110 g). The IGT group also had a significantly higher rate of macrosomia (A kg) than the control group. In that study, maternal age was not a confounding factor as both groups were comparable for age. In our study, the risk of macrosomia in the IGT group remained significantly higher than the NDM group even after exclusion of the older and obese women.

Shoulder dystocia Li et a1 (12) did not find any increase in the rate of

birth trauma in their study. Our study, on the other hand, showed a significantly increased risk of shoul- der dystocia in the IGT group (RR 3.28). This risk remained significantly increased when older women alone were excluded and when both the older and obese women were excluded but statistical

significance was lost when the obese women alone were excluded even though the relative risk remained higher. This is probably due to the low incidence of this condition.

This study clearly shows that impaired glucose tolerance in pregnancy is not a benign condition. I t is definitely associated with increased risks even when maternal age and obesity are excluded as confounding factors.

CONCLUSION The incidence of impaired glucose tolerance in

pregnancy was 8.6% in this study. The state of impaired glucose tolerance per se does

not seem to significantly increase the risk of hypertensive disease in pregnancy which is more likely related to the higher proportion of older and obese women in this group. The rate of Caesarean section for fetal distress/thick meconium-stained liquor is also not significantly increased. However, there seems to be a real increase in the risk of macro- somia, shoulder dystocia and Caesarean section for dystocidno progress in labour with impaired glucose tolerance even when maternal age and obesity are excluded as confounding factors and which is not related to the obstetrician’s knowledge of the diagnosis.

Large randomized controlled trials in which glucose tolerance testing is applied universally and where the investigators and the clinicians are blinded to the diagnosis of impaired glucose tolerance will be needed to conclude irrefutably on whether impaired glucose tolerance really increases the risk of adverse pregnancy outcomes and whether treatment improves these outcomes. Such trials will also be able t o determine the effect of the knowledge of the diagnosis on the rate of obstetric intervention.

We should also bear in mind this reminder from Dornhorst (16) - ‘the diagnosis of gestational diabetes provides a unique opportunity to identify those individuals most at risk of noninsulin dependent diabetes at a time when lifestyle interventions would be expected to reduce overall diabetic morbidity’. This is especially important as diabetes mellitus is not uncommon and is associated with significant multiorgan impairment.

References 1 . Tallarigo L. Giampietro 0. Penno G. Miccoli R. Gregori G .

Navelesi R . Relation of glucose tolerance to complications of pregnancy in non-diabetic women. N Engl J Med 1986; 315: 989-992.

2. Abell DA. Editorial comment: Gestational diabetes. Aust NZ J Obstet Gynaecol 1986; 26: I .

3. Ales KL. Santini DL. Should all pregnant women be screened for gestational glucose intolerance? Lancet 1989; 1 : I 187- I 191,

4. Jarrett RJ. Gestational diabetes: a non-entity? BMJ 1993; 306: 37-38.

Y.Y. TAN AND G.S.H. YEO 255

5 . Ng CSA, Lim LS. Chng KP et al. Combined team management of diabetes mellitus in pregnancy. AM Acad Med Sing 1985; 14: 297-302.

6. Kek LP, Ng CSA. Chng KP et al. Extremes of foetal binhweight for gestation in infants of diabetic mothers. Ann Acad Med Singapore 1985; 14: 303-306.

7. Chan DKL. Ho LY, Tan YY, Liauw PCY. Infants of diabetic mothers - a review of perinatal mortality and morbidity. J Sing Paed Soc 1994; 36: 132-138.

8. Wang KW. Balakrishnan V, Liauw PCY, Chua EKM. Vengadasalam D. Tan YT. Gestational diabetes: What size the problem? Sing Med J 1988; 29: 53-55.

9. Al-Shawaf T, Moghraby S, Akiel A. Does impaired glucose tolerance imply a risk in pregnancy? Br J Obstet Gynaecol 1988; 95: 1036-1041.

10. Roberts RN. Moohan JM. Foo RLK, Harley JMG. Traub AI, Hadden DR. Fetal outcomes in mothers with impaired glucose tolerance in pregnancy. Diabetic Med 1993; 1 0 438-443.

Ausr. h'2 J Obsrer Gwiiecoi 1996; 36: 3: 255

1. Pettitt DJ, Knowler WC, Baird HR, Bennett PH. Gestational diabetes: Infant and maternal complications of pregnancy in relation to third-trimester glucose tolerance in the Pima Indians. Diabetes Care 1980; 3: 458-464.

2. Li DFH, Wong VCW, O'Hoy KMKY, Yeung CY, Ma HK. Is treatment needed for mild impairment of glucose tolerance in pregnancy? Arandomized controlled trial. Br J Obstet Gynaecol 1987; 94: 851-854.

13. Johnstone FD. Nasrat AA, Prescott RJ. The effect of established and gestational diabetes on pregnancy outcome. Br J Obstet Gynaecol 1990 97: 1009-1015.

14. Green JR, Schumacher LB, Pawson IG, Partridge C. Kretchmer N. Influence of maternal body habitus and glucose tolerance on birth weight. Obstet Gynecol 1991; 78: 235-239.

15. Weeks JW, Major CA. de Veciana M, Morgan MA (1994). Gestational diabetes: Does the presence of risk factors influence perinad outcome?Am J Obstet Gynecol 1994; 171: 1003-1007.

16. Domhorst A. Review: Implications of gestational diabetes for the health of the mother. Br J Obstet GynaecoI1994,101: 286-290.

Free Fatty Acids and Gestational Diabetes Mellitus

B. Meyer', D. Calvert' and R. Moses3 Illawarra Area Health Service, New South Wales

Summary: Dyslipidaemia is a common finding in the syndrome of gestational diabetes mellitus (GDM). Free fatty acids (FFA) can cause insulin resistance and have a toxic effect on beta cell function. GDM is a syndrome due to insulin resistance and the purpose of this study was to examine FFA levels in pregnancy. For this purpose samples were taken from 44 consecutive women with GDM and 36 consecutive controls. The women with GDM had similar levels of insulin, triglycerides and total cholesterol to the controls but a significant elevation in FFA; 0.70 (0.33) mmoYL versus 0.29 (0.12) m m o a (p<O.O001). A raised level of FFA may be responsible for the insulin resistance of GDM and may assist in the diagnosis.

Gestational diabetes mellitus (GDM) is carbo- hydrate intolerance of variable severity with onset or first recognition during the current pregnancy (1). GDM may complicate between 5-7% (2,3) of pregnancies in Australia and is due to the failure of maternal beta cell secretion to overcome the normal insulin resistance of pregnancy (4).

While all types of diabetes are invariably defined after glucose tolerance testing, and usually managed on the basis of glucose measurements, diabetes is a complex metabolic disorder. Diabetes in pregnancy,

1. Lecturer, Medical Research Unit. 2. Head, Medical Research Unit. 3. Clinical Director of Diabetes Services. Address for correspondence: Dr R.G. Moses, 4/393 Crown Street, Wollongong, New South Wales, 2500.

like other forms of diabetes, can be associated with abnormalities of lipid and protein metabolism (5). Tri- glyceride levels increase during pregnancy, initially due to an increased hepatic production (6) and later due to a resistance to peripheral uptake (7). Dyslipid- aemia may be important in pregnancy and prognostic of certain outcomes. An early elevation of triglycer- ides may be predictive of GDM (8) and women who develop preeclampsia have a preceding rise in both triglycerides and free fatty acids (FFA) (9). Insulin resistance can be caused by increased levels of FFA (10) and increased levels of FFA may have a toxic effect on beta cell function (1 1).

The glucose tolerance test has limitations for diagnosing GDM (12) and women with excellent glucose control during pregnancy may still have a large or macmmic fetus, a clear indication that