CMAJ Research

Effect of maternal weight, adipokines, glucose intolerance

and lipids on infant birth weight among women without
gestational diabetes mellitus

Ravi Retnakaran MD, Chang Ye MSc, Anthony J.G. Hanley PhD, Philip W. Connelly PhD, Mathew Sermer MD,
Bernard Zinman MD, Jill K. Hamilton MD

See related commentary by Ryan on page 1341 and at www.cmaj.ca/lookup/doi/10.1503/cmaj.120682

Abstract Competing interests: None

Background: The delivery of excess maternal linear regression analysis, positive determi- declared.
nutrients to the fetus is known to increase the nants of birth weight were length of gesta- This article has been peer
risk of macrosomia, even among infants of wo- tion, male infant, weight gain during preg- reviewed.
men without gestational diabetes mellitus. nancy up to the time of the oral glucose
Correspondence to:
With the current obesity epidemic, maternal tolerance test, body mass index (BMI) before Jill K. Hamilton,
adiposity and its associated effects on circulat- pregnancy and impaired glucose tolerance in [email protected]
ing adipokines and inflammatory proteins may pregnancy. Leptin, adiponectin and C-reactive
CMAJ 2012. DOI:10.1503
now have a greater impact on fetal growth. We protein levels were each negatively associated
/cmaj.111154
sought to evaluate the independent effects of with birth weight. On logistic regression analy-
maternal glycemia, lipids, obesity, adipokines sis, the significant metabolic predictors of hav-
and inflammation on infant birth weight. ing a large-for-gestational-age infant were
BMI before pregnancy (odds ratio [OR] 1.16,
Methods: We included 472 women who under- 95% confidence interval [CI] 1.05–1.27, per
went an oral glucose tolerance test in late 1 kg/m2 increase), weight gain during preg-
pregnancy and were found not to have gesta- nancy up to the time of the oral glucose toler-
tional diabetes; 104 (22.0%) had gestational ance test (OR 1.12, 95% CI 1.05–1.19, per 1 kg
impaired glucose tolerance. We also measured increase) and leptin level (OR 0.50, 95% CI
fasting levels of insulin, low- and high-density 0.30–0.82, per 1 standard deviation change).
lipoprotein cholesterol, triglycerides, leptin,
adiponectin and C-reactive protein. Obstetric Interpretation: Among women without gesta-
outcomes were assessed at delivery. tional diabetes, maternal adiposity and leptin
levels were the strongest metabolic determi-
Results: The mean birth weight was 3481 g nants of having a large-for-gestational-age
(standard deviation 493 g); 68 of the infants infant rather than glucose intolerance and
were large for gestational age. On multiple lipid levels.

I
n 1952, Jørgen Pedersen proposed that betes or gestational diabetes, modern clinical
delivery of excess maternal glucose to the practice focuses on normalizing blood glucose
fetus may be responsible for the increased levels to reduce the risk of fetal hyperglycemia
risk of macrosomia among infants of women and hence the risk of fetal macrosomia and its
with diabetes during pregnancy.1 He postulated associated adverse clinical outcomes (e.g., shoul-
that maternal hyperglycemia leads to fetal hyper- der dystocia, birth injury, need for cesarean
glycemia, which in turn stimulates insulin secre- delivery).
tion in the fetus, the anabolic effects of which It is now recognized that the association be-
result in excessive fetal growth. Since its intro- tween maternal nutrients and fetal growth is not
duction, the Pedersen hypothesis has been fur- restricted solely to women with diabetes. Several
ther extended by other investigators and accepted studies have shown associations linking maternal
as the pathophysiologic basis for increased risk blood glucose and triglyceride levels with infant
of macrosomia among infants of women with birth weight among women without gestational
diabetes during pregnancy.2,3 Accordingly, for diabetes.4–7 This awareness has led to recent rec-
pregnant women with either pre-existing dia- ommendations to lower the diagnostic thresholds

© 2012 Canadian Medical Association or its licensors CMAJ, September 4, 2012, 184(12) 1353
Research

for gestational diabetes on glucose tolerance test- the groups for which Canadian-based ethnicity-
ing in pregnancy, to optimize the detection of specific centiles of birth weight have been estab-
women who may be at risk of having a large-for- lished.14,15 Lastly, the analysis was restricted to
gestational-age infant.8 women with term (37–41 weeks’ gestation inclu-
Another important factor relevant to the risk sive) singleton pregnancies, because both prema-
of macrosomia is maternal adiposity.9 Indeed, turity and twin pregnancy can affect infant
the past decade has seen a marked increase in growth.
the prevalence of pre-existing obesity among
pregnant women.10 In the context of the current Biochemical analysis
obesity epidemic, we hypothesized that, in Fasting serum samples for biochemical analysis
women without gestational diabetes, maternal were obtained at the time of the oral glucose
adiposity and its associated effects on circulat- tolerance test. Fasting insulin levels were mea-
ing levels of adipokines (e.g., adiponectin and sured with the Roche Elecsys 1010 immuno-
leptin) and inflammatory proteins (C-reactive assay analyzer and electrochemiluminescence
protein) may now have a greater impact than immunoassay kit (Roche Diagnostics, Laval,
glucose and lipid levels on fetal growth. We Quebec).
conducted this study to evaluate the independent Levels of total cholesterol, high-density
effects of maternal glycemia, lipid levels, obe- lipoprotein (HDL) cholesterol and triglycerides
sity, adipokine levels and inflammation on the were measured with the Roche Cobas 6000 c 501
infant birth weight in a cohort of women with- analyzer (Roche Diagnostics, Laval). Levels of
out gestational diabetes. low-density lipoprotein (LDL) cholesterol were
estimated with use of the Friedwald formula.
Methods Apolipoprotein B and A-I levels were measured
with the BN ProSpec System (Siemens Health-
We conducted this analysis within an ongoing care Diagnostics, Mississauga, Ontario).
prospective observational cohort study, in which Total adiponectin levels were measured with
women are recruited at the time of antepartum an enzyme-linked immunosorbent assay (Milli-
screening for gestational diabetes and followed pore Corporation, Linco Research Inc., St.
into the postpartum period. The protocol for the Charles, Missouri). High-sensitivity C-reactive
ongoing study has been described in detail previ- protein levels were measured by means of end-
ously.11,12 In brief, following recruitment late in point nephelometry with use of the BN
the second or early in the third trimester, all par- ProSpec Analyzer and N high-sensitivity C-
ticipants undergo a three-hour 100-g oral glucose reactive protein reagent (Dade–Behring, Mis-
tolerance test. Based on the results, the women sissauga). Leptin levels were measured by
are classified as having gestational diabetes means of enzyme-linked immunosorbent assay
(defined as two or more glucose values above the (model EZHL-80SK, Millipore Corporation,
National Diabetes Data Group threshold),13 ges- Linco Research Inc., St. Charles). The intra-
tational impaired glucose tolerance (defined as and intervariation coefficients for these assays
only one value above the threshold) or normal have been reported previously.16
glucose tolerance.11 Serum and plasma samples
are obtained at the time of this test for assess- Obstetric outcomes
ment of biochemical factors. Obstetric outcomes At delivery, data were collected on the length of
are assessed at delivery. The protocol has been gestation, the infant’s sex, birth weight and
approved by the Mount Sinai Hospital Research Apgar scores, and the mode of delivery. A large-
Ethics Board, and all participants have provided for-gestational-age infant was defined as one
written informed consent. whose sex-specific birth weight for gestational
For the current analysis, we included partici- age was above the 90th percentile of Canadian
pants who were identified as having either gesta- fetal growth curves for the relevant ethnic group
tional impaired glucose tolerance or normal glu- (white, Asian or South Asian).14,15 Macrosomia
cose tolerance (including those with an abnormal was defined as a birth weight of 4000 g or
screening glucose challenge test). We excluded higher.
women with gestational diabetes because the
glucose-lowering treatment (diet or insulin) that Statistical analysis
they would receive to reduce the risk of exces- We assessed maternal glycemia according to glu-
sive fetal growth would obscure the associations cose tolerance status and area-under-the-glucose-
between maternal factors and birth weight. The curve on the oral glucose tolerance test. We
study population was further restricted to white, tested continuous variables for normality of dis-
Asian and South Asian women because these are tribution, applying natural log-transformations of

1354 CMAJ, September 4, 2012, 184(12)

 Research

skewed variables where necessary. We stratified insulin and triglycerides, and lower levels of
the study population into tertiles based on infant adiponectin.
birth weight. We compared continuous variables At delivery, the top tertile had the longest
using analysis of variance and categorical vari- length of gestation and the largest proportion of
ables using the χ2 or Fisher exact test. cesarean deliveries. Although the proportion of
Multiple linear regression analysis of infant infants with an Apgar score below 7 at one
birth weight was performed, with adjustment for minute differed significantly between the three
the following covariates: length of gestation, groups, the proportion with a score below 7 at
infant sex, maternal demographic factors (age, five minutes did not. As expected, there were no
ethnicity, family history of diabetes), smoking instances of macrosomia or large-for-gestational-
status, anthropometric measures (body mass age infants in the lower two tertiles.
index [BMI] before pregnancy, and weight gain
during pregnancy up to the time of the oral glu- Maternal factors associated with infant
cose tolerance test), glucose tolerance status, birth weight
lipids (LDL, HDL and triglycerides), insulin, The unadjusted associations between maternal and
adipokines (adiponectin, leptin) and inflamma- fetal factors and birth weight are shown in Table 2.
tory proteins (C-reactive protein). Length of gestation, male infant, BMI before preg-
Logistic regression analysis of the dependent nancy and weight gain during pregnancy up to the
variable (large-for-gestational-age infant) was time of the oral glucose tolerance test were all
performed with use of the same covariates as in associated with higher birth weight. South Asian
the multiple linear regression analysis, except for ethnicity and adiponectin level were associated
length of gestation and infant sex because both with lower birth weight. After adjustment for co-
of these variables are considered when defining variates, positive determinants of birth weight were
large-for-gestational-age infants. length of gestation, male infant, BMI before preg-
Sensitivity analyses were performed in which nancy, weight gain during pregnancy up to the
the homeostasis model assessment of β-cell time of the oral glucose tolerance test and gesta-
function and insulin resistance was used in place tional impaired glucose tolerance. Leptin, adipo-
of fasting insulin, and the area-under-the glu- nectin and C-reactive protein were each negatively
cose-curve was used in place of gestational associated with birth weight. This model explained
impaired glucose tolerance. These sensitivity 25.9% of the variance in birth weight.
analyses were performed to assess robustness of On logistic regression analysis, the significant
the findings upon adjustment for β-cell function, predictors of having a large-for-gestational-age
insulin resistance and a continuous measure of infant, after adjustment for covariates, were
glycemia. Asian ethnicity (v. white) (OR 3.02, 95% CI
All analyses were performed with the use of 1.25–7.27), BMI before pregnancy (OR 1.16,
Statistical Analysis System 9.2 (SAS Institute, 95% CI 1.05–1.27, per 1 kg/m2 increase), weight
Cary, North Carolina). gain during pregnancy up to the time of the oral
glucose tolerance test (OR 1.12, 95% CI 1.05–
Results 1.19, per 1 kg increase) and leptin level (OR
0.50, 95% CI 0.30–0.82, per 1 SD change) (Fig-
Antepartum characteristics and obstetric ure 1, top panel). Because Asian women were
outcomes found to have a greater risk of ethnicity-specific
Of the 472 women included in our study, 368 large-for-gestational-age infants compared with
had normal glucose tolerance and 104 had gesta- white women, we repeated the logistic regres-
tional impaired glucose tolerance. Overall, the sion analysis and included only the white women
infants had a mean birth weight of 3481 g (stan- (n = 388). In this analysis, BMI before preg-
dard deviation [SD] 493 g), and 68 were large nancy (OR 1.13, 95% CI 1.02–1.24, per 1 kg/m2
for gestational age at delivery. increase), weight gain during pregnancy up to
Table 1 shows the antepartum characteristics the time of the oral glucose tolerance test (OR
and obstetric outcomes stratified by tertile of 1.10, 95% CI 1.03–1.18, per 1 kg increase) and
infant birth weight. The three groups differed in leptin level (OR 0.52, 95% CI 0.29–0.91, per
ethnicity, with the lowest tertile having the 1 SD change) again emerged as significant pre-
greatest proportion of Asian and South Asian dictors of large-for-gestational-age infants; being
women. As expected, BMI before pregnancy a former smoker (v. never smoking) also reached
differed across the groups and was highest in significance (OR 2.00, 95% CI 1.02–3.93) (Fig-
the top tertile of birth weight. Consistent with ure 1, bottom panel).
this difference in BMI, the top tertile also The sensitivity analyses identified the same
exhibited higher levels of fasting glucose, independent predictors on multiple linear regres-

CMAJ, September 4, 2012, 184(12) 1355

Research

Table 1: Characteristics and pregnancy outcomes of 472 women without gestational diabetes, by tertile of infant birth weight

Tertile of infant birth weight; no. (% of women)*

Lowest tertile Middle tertile Highest tertile
[2020–3260 g] [3260–3670 g] [3670–5700 g]
Variable n = 156 n = 157 n = 159 p value†

At oral glucose tolerance test

Age, yr, mean (SD) 33.6 (4.0) 34.5 (4.3) 33.6 (4.0) 0.08
No. of weeks of gestation, median (IQR) 30 (29–32) 30 (28–32) 30 (28–31) 0.06
Ethnicity 0.009
White 118 (75.6) 134 (85.4) 136 (85.5)
Asian 23 (14.7) 16 (10.2) 21 (13.2)
South Asian 15 (9.6) 7 (4.5) 2 (1.3)
Family history of diabetes 87 (55.8) 81 (51.6) 90 (56.6) 0.6
Parity 0.08
0 94 (60.3) 78 (49.7) 79 (49.7)
1 52 (33.3) 61 (38.9) 56 (35.2)
>1 10 (6.4) 18 (11.5) 24 (15.1)
Previous gestational diabetes 5 (3.2) 8 (5.1) 11 (6.9) 0.3
Smoking status 0.9
Never smoked 103 (66.0) 109 (69.4) 105 (66.0)
Former smoker 49 (31.4) 44 (28.0) 51 (32.1)
Current smoker 4 (2.6) 4 (2.6) 3 (1.9)
BMI before pregnancy, kg/m2, median (IQR) 22.6 (20.7–25.4) 22.6 (20.8–25.8) 23.6 (22.3–27.4) < 0.001
Weight gain during pregnancy up to the oral 10.8 (5.4) 11.6 (4.9) 11.6 (5.7) 0.3
glucose tolerance test, kg, mean (SD)
Weight gain per week of gestation, kg/wk, mean (SD) 0.36 (0.18) 0.38 (0.16) 0.39 (0.18) 0.3
Fasting glucose, mmol/L, mean (SD) 4.4 (0.4) 4.4 (0.4) 4.6 (0.5) 0.001
Area-under-glucose-curve, mean (SD) 21.1 (2.6) 21.4 (3.0) 21.5 (2.8) 0.5
Glucose tolerance in pregnancy 0.08
Normal 131 (84.0) 117 (74.5) 120 (75.5)
Impaired 25 (16.0) 40 (25.5) 39 (24.5)
Fasting insulin, pmol/L, median (IQR) 57 (39–79) 53 (36–71) 65 (42–94) < 0.001
Total cholesterol, mmol/L, mean (SD) 6.48 (1.25) 6.55 (1.23) 6.39 (1.15) 0.5
LDL cholesterol, mmol/L, mean (SD) 3.72 (1.17) 3.72 (1.12) 3.6 (1.04) 0.5
HDL cholesterol, mmol/L, mean (SD) 1.73 (0.36) 1.72 (0.37) 1.66 (0.34) 0.2
Triglycerides, mmol/L, mean (SD) 2.25 (0.72) 2.46 (0.75) 2.49 (0.66) 0.006
Apolipoprotein B, g/L, mean (SD) 1.28 (0.31) 1.30 (0.30) 1.29 (0.28) 0.7
Apolipoprotein A-I, mmol/L, mean (SD) 2.10 (0.28) 2.10 (0.32) 2.07 (0.26) 0.5
C-reactive protein, mg/L, median (IQR) 4.5 (2.5–7.6) 4.1 (2.0–7.5) 3.8 (2.3–7.1) 0.7
Adiponectin, µg/mL, median (IQR) 8.2 (6.3–10.0) 8.6 (6.7–10.4) 7.5 (6.0–9.2) 0.02
Leptin, ng/mL, median (IQR) 34.5 (25.3–46.4) 31.7 (19.8–45.8) 33.4 (23.1–46.2) 0.7
At delivery
Length of gestation, wk, mean (SD) 38.6 (1.1) 39.2 (1.0) 39.6 (1.1) < 0.001
Infant sex, male/female, % 40/60 50/50 52/48 0.07
Infant with Apgar score < 7 at 1 min 12 (7.8) 2 (1.3) 10 (6.3) 0.02
Infant with Apgar score < 7 at 5 min 0 0 2 (1.3) 0.1
Cesarean delivery 49 (31.4) 49 (31.2) 75 (47.2) 0.003
Infant with macrosomia 0 0 61 (38.4) < 0.001
Large-for-gestational-age infant 0 0 68 (42.8) < 0.001
Note: BMI = body mass index, HDL = high-density lipoprotein, IQR = interquartile range, LDL = low-density lipoprotein, SD = standard deviation.
*Unless specified otherwise.
†Values refer to overall differences across the groups, as determined by analysis of variance for continuous variables, or χ test or Fisher exact test for categorical variables.
2

1356 CMAJ, September 4, 2012, 184(12)

 Research

sion and logistic regression as those in Table 2 study pertains to the relative contributions of these
and Figure 1 (top panel), respectively. obesity-related circulating factors. Specifically, we
found leptin to be a significant negative predictor
Interpretation of both birth weight and large-for-gestational-age
infant after adjustment for covariates, including
Our study yielded three key findings. First, the BMI before pregnancy and weight gain during
strongest determinant of infant birth weight and pregnancy. An inverse relation between leptin level
of having a large-for-gestational-age infant was and infant birth weight has been observed in some,
maternal adiposity, with BMI before pregnancy though not all, previous studies.21–24 Of note, Ver-
and weight gain during pregnancy emerging as haeghe and colleagues found that, compared with
the strongest predictors for both outcomes in the glucose, insulin, adiponectin and tumour necrosis
fully adjusted analyses. This finding is consistent factor-α, leptin was the mediator that was most
with those of previous studies, which suggests strongly (and negatively) associated with birth
the central role of maternal weight as a determi- weight.23 Placental dysfunction has been suggested
nant of birth weight.17–19 Although the underlying as a unifying mechanism underlying both fetal
mechanisms remain unclear, shared genetic fac- growth restriction and elevated maternal leptin lev-
tors may be relevant to both maternal adiposity els (through increased placental production of lep-
and fetal growth. tin), although this model remains speculative.24
Obesity is associated with pathologic sequelae, Like leptin, C-reactive protein also emerged as
including chronic inflammation (increased C- a negative determinant of infant birth weight in
reactive protein level) and adipocyte dysregulation our multiple linear regression analysis. In the
(increased leptin and low adiponectin levels).20 In Hyperglycemia and Adverse Pregnancy Outcome
this context, the second key observation of our Study, an inverse relation between C-reactive pro-

Table 2: Change in infant birth weight in relation to maternal and fetal factors*

Change in infant birth weight, g (95% CI)

Variable Crude Adjusted†

Length of gestation, wk 166.65 (129.99 to 203.32) 164.85 (127.50 to 202.22)

Male infant 115.76 (26.97 to 204.54) 139.33 (55.55 to 223.11)
Age, yr 2.24 (–8.60 to 13.08) –0.26 (–10.49 to 9.97)
Ethnicity (v. white)
Asian –102.78 (–236.45 to 30.89) –28.48 (–169.09 to 112.13)
South Asian –322.67 (–523.73 to –121.62) –165.12 (–375.24 to 44.99)
Family history of diabetes 27.66 (–61.96 to 117.28) 15.31 (–69.47 to 100.09)
Smoking status (v. never smoked)
Former 52.64 (–44.17 to 149.46) 45.84 (–44.90 to 136.58)
Current –79.35 (–375.10 to 216.40) –187.36 (–464.34 to 89.62)
BMI before pregnancy, kg/m2 15.22 (5.66 to 24.78) 22.45 (9.66 to 35.25)
Weight gain during pregnancy up to
oral glucose tolerance test, kg 9.49 (1.11 to 17.87) 19.15 (9.93 to 28.37)
Impaired glucose tolerance in
pregnancy (v. normal glucose tolerance) 100.29 (–7.01 to 207.59) 120.37 (19.90 to 220.84)
Fasting insulin, pmol/L 0.63 (–0.06 to 1.32) 0.16 (–0.53 to 0.85)
LDL cholesterol, mmol/L –15.22 (–55.49 to 25.05) –6.79 (–46.98 to 33.39)
HDL cholesterol, mmol/L –120.54 (–244.42 to 3.35) –57.16 (–189.42 to 75.09)
Triglycerides, mmol/L 61.11 (–1.18 to 123.40) –1.59 (–70.67 to 67.49)
C-reactive protein, mg/L –3.68 (–10.98 to 3.63) –7.67 (–14.88 to –0.47)
Adiponectin, µg/mL –21.53 (–37.67 to –5.38) –19.35 (–36.62 to –2.07)
Leptin, per ng/mL –0.81 (–2.79 to 1.16) –3.92 (–6.23 to –1.60)

Note: BMI = body mass index, CI = confidence interval, HDL = high-density lipoprotein, LDL = low-density lipoprotein.
*Change in infant birth weight per unit change in indicated variable. For example, infant birth weight increased by 164.85 g
per additional week of gestation after adjustment for the other variables.
†Adjusted for all other variables listed.

CMAJ, September 4, 2012, 184(12) 1357

Research

tein level and birth weight was seen after adjust- protein in the setting of obesity (which itself is the
ment for covariates, including maternal BMI.25 primary determinant of increased birth weight),
The negative associations of leptin and C-reactive both of these proteins were found to be indepen-
protein with birth weight became apparent in our dently associated with decreased birth weight.
study only after adjustment for covariates, includ- Although the mechanisms are unclear, our data
ing BMI before pregnancy. Thus, despite the ele- suggest that these factors may play a role in atten-
vated circulating levels of leptin and C-reactive uating the pro-macrosomia effects of maternal

Unadjusted* odds Adjusted* odds

Variable* ratio (95% CI) ratio (95% CI) Decreased Increased
risk risk
All women
Age, yr 1.02 (0.96–1.08) 1.00 (0.93–1.07)
Asian ethnicity (v. white) 1.97 (1.01–3.83) 3.02 (1.25–7.27)
South Asian ethnicity (v. white) 0.59 (0.13–2.57) 1.45 (0.28–7.44)
Family history of diabetes mellitus 1.41 (0.83–2.38) 1.62 (0.88–2.98)
Former smoker (v. never smoked) 1.53 (0.89–2.62) 1.82 (0.99–3.37)
Current smoker (v. never smoked) 1.54 (0.32–7.38) 1.75 (0.29–10.69)
BMI before pregnancy, kg/m2 1.04 (0.99–1.10) 1.16 (1.05–1.27)
Weight gain during pregnancy 1.05 (1.00–1.10) 1.12 (1.05–1.19)
up to oral glucose tolerance test, kg
Impaired glucose tolerance 1.59 (0.89–2.82) 1.36 (0.71–2.61)
in pregnancy
Fasting insulin, pmol/L 1.09 (0.89–1.33) 0.97 (0.70–1.34)
LDL cholesterol, mmol/L 0.80 (0.61–1.05) 0.98 (0.72–1.34)
HDL cholesterol, mmol/L 0.89 (0.69–1.15) 0.99 (0.70–1.39)
Triglycerides, mmol/L 1.26 (0.98–1.62) 0.98 (0.70–1.38)
C-reactive protein, mg/L 0.79 (0.57–1.10) 0.75 (0.50–1.12)
Adiponectin, µg/mL 0.68 (0.50–0.92) 0.79 (0.55–1.13)
Leptin, ng/mL 0.85 (0.63–1.13) 0.50 (0.30–0.82)

0.25 0.5 1 2 4 8 16
White women
Age, yr 1.02 (0.95–1.09) 1.01 (0.93–1.10)
Family history of diabetes mellitus 1.56 (0.85–2.85) 1.65 (0.83–3.28)
Former smoker (v. never smoked) 1.75 (0.96–3.20) 2.00 (1.02–3.93)
Current smoker (v. never smoked) 0.81 (0.37–8.82) 1.79 (0.29–10.83)
BMI before pregnancy, kg/m2 1.04 (0.99–1.10) 1.13 (1.02–1.24)
Weight gain during pregnancy 1.06 (1.00–1.11) 1.10 (1.03–1.18)
up to oral glucose tolerance test, kg
Impaired glucose tolerance 1.34 (0.68–2.65) 1.34 (0.62–2.90)
in pregnancy
Fasting insulin, pmol/L 1.06 (0.85–1.32) 0.90 (0.54–1.48)
LDL cholesterol, mmol/L 0.85 (0.62–1.16) 0.98 (0.69–1.38)
HDL cholesterol, mmol/L 0.82 (0.60–1.10) 1.03 (0.69–1.52)
Triglycerides, mmol/L 1.33 (1.00–1.77) 1.07 (0.73–1.58)
C-reactive protein, mg/L 0.80 (0.55–1.16) 0.69 (0.43–1.12)
Adiponectin, µg/mL 0.64 (0.45–0.92) 0.70 (0.46–1.08)
Leptin, ng/mL 0.88 (0.63–1.23) 0.52 (0.29–0.91)

0.25 0.5 1 2 4 8 16
Adjusted odds ratio (95% CI)

Figure 1: Independent predictors of having a large-for-gestational-age infant among women without gestational diabetes, in the entire
study population (top panel) and among white women only (bottom panel). An odds ratio greater than 1.00 is associated with an
increased risk of having a large-for-gestational-age infant. Odds ratios for continuous variables are expressed per 1 standard deviation
change in the indicated variable, except for age (expressed per year increase), body mass index (BMI) before pregnancy (expressed per
1 kg/m2 increase) and weight gain during pregnancy (expressed per 1 kg increase). For example, the adjusted odds ratio for having a
large-for-gestational-age infant increased by 1.16 per 1 kg/m2 increase in BMI before pregnancy. *Each variable was adjusted for all
others in the list. CI = confidence interval, HDL = high-density lipoprotein, LDL = low-density lipoprotein.

1358 CMAJ, September 4, 2012, 184(12)

 Research

adiposity. Overall, coupled with the inverse rela- Strengths and limitations
tion between total adiponectin level and birth A major strength of our analysis is that it evalu-
weight, these data suggest a model in which fetal ated simultaneously the effect of several mater-
growth is influenced both positively and nega- nal metabolic factors, including glucose toler-
tively by complex interplay between maternal adi- ance, adiposity, and fasting levels of insulin,
posity and its associated circulating factors. lipids, adipokines and inflammatory proteins.
The third key finding from our study is that, This approach enabled the assessment of the fac-
although gestational impaired glucose tolerance tors’ independent associations with infant birth
was an independent predictor of birth weight in weight after adjustment for one another.
the multiple linear regression analysis, its impact One limitation of our study is that we did not
was relatively modest compared with that of BMI evaluate the high-molecular-weight form of
before pregnancy, weight gain during pregnancy adiponectin. Although total and high-molecular-
up to the time the oral glucose tolerance test and weight adiponectin are generally highly corre-
the leptin level. Moreover, unlike the latter three lated, the high-molecular-weight form may be
factors, gestational impaired glucose tolerance the specific maternal mediator affecting fetal
was not a significant independent predictor of growth.28 Another limitation is that weight before
having a large-for-gestational-age infant. Simi- pregnancy was determined by patient recall and
larly, none of the lipid measures was indepen- hence may have been subject to bias. In addition,
dently associated with birth weight or large-for- weight gain during pregnancy was determined
gestational-age infant. These data suggest that only up to the time of the oral glucose tolerance
maternal weight and its associated circulating test and not across the total length of gestation.
factors have a greater impact on infant birth However, the impact of this limitation should be
weight than do mild glucose intolerance and lipid modest, because changes in maternal weight in
levels in women without gestational diabetes. the first and second trimesters have been shown
This concept is consistent with findings from to have a much greater influence on birth weight
two recent studies addressing the relative effects than changes in the third trimester.29
of maternal obesity and glycemia on fetal
growth. In a re-analysis of data from the Hyper- Conclusion
glycemia and Adverse Pregnancy Outcome Among women without gestational diabetes, the
Study, Ryan recently showed that maternal BMI strongest metabolic determinant of macrosomia
had a greater impact on the odds of having a was maternal adiposity, as reflected by BMI
large-for-gestational-age infant than did maternal before pregnancy and weight gain during preg-
glucose in all but the highest category of nancy up to the time of the oral glucose tolerance
glycemia. 26 In addition, in a study involving test. Furthermore, after adjustment for these
Spanish women, Ricart and colleagues found attributes, weight-related circulating factors (par-
that BMI before pregnancy commanded a much ticularly leptin) emerged as negative independent
higher population-attributable risk for macroso- predictors of having a large-for-gestational-age
mia than did gestational hyperglycemia.17 Our infant. In contrast, glucose intolerance and lipid
study further extends this discussion by showing levels had a modest effect in this cohort.
that both BMI before pregnancy and weight gain In the context of the current obesity epi-
during pregnancy had a greater independent demic, these data support the importance of tar-
impact on the risk of excessive fetal growth than geting healthy body weight in young women as
did glycemia in women without gestational a strategy for reducing the risk of excessive fetal
diabetes. growth and infant macrosomia. Furthermore,
This emerging concept regarding the relative these findings suggest that, in the care of over-
importance of maternal obesity versus glycemia weight or obese women in pregnancy, closer
may hold implications for the current debate monitoring of weight gain during pregnancy
regarding the lowering of diagnostic thresholds may be warranted.
for gestational diabetes on antepartum glucose
tolerance testing. 26,27 Specifically, among the References
additional women who will be identified as hav- 1. Pedersen J. Diabetes and pregnancy: blood sugar of newborn
ing gestational diabetes owing to this change, infants [dissertation]. Copenhagen (Denmark): Danish Science
Press; 1952. p.230.
their mild glucose intolerance may play a com- 2. Freinkel N. Banting lecture 1980. Of pregnancy and progeny.
paratively modest role in determining risk of Diabetes 1980;29:1023-35.
3. Lindsay RS. Many HAPO returns: maternal glycemia and
macrosomia. In this context, it may instead be neonatal adiposity: new insights from the Hyperglycemia and
more prudent to target maternal obesity, the Adverse Pregnancy Outcomes (HAPO) study. Diabetes 2009;58:
302-3.
prevalence of which has risen dramatically in 4. Sermer M, Naylor CD, Gare DJ, et al. Impact of increasing car-
recent years.10 bohydrate intolerance on maternal–fetal outcomes in 3637 wo-

CMAJ, September 4, 2012, 184(12) 1359

Research

men without gestational diabetes. The Toronto Tri-Hospital Ges- birthweight: Can insulin or adipokines do better? Metabolism
tational Diabetes Project. Am J Obstet Gynecol 1995;173:146-56. 2006;55:339-44.
5. Metzger BE, Lowe LP, Dyer AR, et al.; HAPO Study Coopera- 24. Pighetti M, Tommaselli GA, D’Elia A, et al. Maternal serum and
tive Research Group. Hyperglycemia and adverse pregnancy umbilical cord blood leptin concentrations with fetal growth
outcomes. N Engl J Med 2008;358:1991-2002. restriction. Obstet Gynecol 2003;102:535-43.
6. Knopp RH, Magee MS, Walden CE, et al. Prediction of infant 25. Lowe LP, Metzger BE, Lowe WL Jr, et al.; HAPO Study Coop-
birth weight by GDM screening tests. Importance of plasma erative Research Group. Inflammatory mediators and glucose in
triglyceride. Diabetes Care 1992;15:1605-13. pregnancy: results from a subset of the Hyperglycemia and
7. Di Cianni G, Miccoli R, Volpe L, et al. Maternal triglyceride lev- Adverse Pregnancy Outcome (HAPO) Study. J Clin Endocrinol
els and newborn weight in pregnant women with normal glucose Metab 2010;95:5427-34.
tolerance. Diabet Med 2005;22:21-5. 26. Ryan EA. Diagnosing gestational diabetes. Diabetologia 2011;
8. Metzger BE, Gabbe SG, Persson B, et al. International Associa- 54:480-6.
tion of Diabetes and Pregnancy Study Groups recommendations 27. Long H. Diagnosing gestational diabetes: Can expert opinions
on the diagnosis and classification of hyperglycemia in preg- replace scientific evidence? Diabetologia 2011;54:2211-3.
nancy. Diabetes Care 2010;33:676-82. 28. Ong GK, Hamilton JK, Sermer M, et al. Maternal serum
9. Catalano PM, Hauguel-De Mouzon S. Is it time to revisit the adiponectin and infant birthweight: the role of adiponectin iso-
Pedersen hypothesis in the face of the obesity epidemic? Am J form distribution. Clin Endocrinol (Oxf) 2007;67:108-14.
Obstet Gynecol 2011;204:479-87. 29. Brown JE, Murtaugh MA, Jacobs DR, et al. Variation in new-
10. Kanagalingam MG, Forouhi NG, Greer IA, et al. Changes in born size according to pregnancy weight change by trimester.
booking body mass index over a decade: retrospective analysis Am J Clin Nutr 2002;76:205-9.
from a Glasgow Maternity Hospital. BJOG 2005;112:1431-3.
11. Retnakaran R, Qi Y, Sermer M, et al. Glucose intolerance in
Affiliations: From the Leadership Sinai Centre for Diabetes
pregnancy and future risk of pre-diabetes or diabetes. Diabetes
Care 2008;31:2026-31. (Retnakaran, Ye, Hanley, Zinman), the Division of Obstetrics
12. Retnakaran R, Qi Y, Connelly PW, et al. Low adiponectin con- and Gynecology (Sermer) and the Samuel Lunenfeld
centration during pregnancy predicts postpartum insulin resis- Research Institute (Zinman), Mount Sinai Hospital; the Divi-
tance, beta-cell dysfunction and fasting glycaemia. Diabetologia sion of Endocrinology (Retnakaran, Hanley, Connelly, Zin-
2010;53:268-76. man, Hamilton) and the Department of Nutritional Sciences
13. National Diabetes Data Group. Classification and diagnosis of (Hanley), University of Toronto; the Keenan Research Centre
diabetes mellitus and other categories of glucose intolerance. (Connelly), Li Ka Shing Knowledge Institute, St. Michael’s
Diabetes 1979;28:1039-57.
Hospital; and the Division of Endocrinology (Hamilton),
14. Kramer MS, Platt RW, Wen SW, et al.; Fetal/Infant Health Study
Group of the Canadian Perinatal Surveillance System. A new Hospital for Sick Children, Toronto, Ont.
and improved population-based Canadian reference for birth-
Contributors: Ravi Retnakaran designed the analysis plan,
weight for gestational age. Pediatrics 2001;108:E35-41.
15. Kierans WJ, Kramer MS. Wilkins R, et al. Charting birth out- researched the data and wrote the manuscript. Chang Ye
come in British Columbia: determinants of optimal health and performed the statistical analyses. Ravi Retnakaran,
ultimate risk — an expansion and update. Victoria (BC): British Anthony Hanley, Philip Connelly, Mathew Sermer, Bernard
Columbia Vital Statistics Agency; 2003. Available: www.vs.gov Zinman and Jill Hamilton were involved in the design and
.bc.ca/stats/features/index.html (accessed 2011 Mar. 17). implementation of the overall study. All of the authors con-
16. Retnakaran R, Qi Y, Sermer M, et al. The postpartum cardiovas- tributed to critical revision of the manuscript and approved
cular risk factor profile of women with isolated hyperglycemia at the final version submitted for publication.
1-hour on the oral glucose tolerance test in pregnancy. Nutr
Metab Cardiovasc Dis 2011;21:706-12. Funding: This study was supported by operating grants from
17. Ricart W, López J, Mozas J, et al.; Spanish Group for the Study the Canadian Institutes of Health Research (CIHR) (grant
of the Impact of Carpenter and Coustan GDM Thresholds. Body
nos. MOP-89831 and MOP-84206). The study sponsor had
mass index has a greater impact on pregnancy outcomes than
gestational hyperglycemia. Diabetologia 2005;48:1736-42. no role in the design of the study, the collection, analysis or
18. HAPO Study Cooperative Research Group. Hyperglycaemia and interpretation of data, the writing of the report or the decision
Adverse Pregnancy Outcome (HAPO) Study: associations with to submit the article for publication.
maternal body mass index. BJOG 2010;117:575-84.
19. Yu CK, Teoh T, Robinson S. Obesity in pregnancy. BJOG 2006; Acknowledgements: The authors thank Mount Sinai Hospi-
113:1117-25. tal’s Department of Pathology and Laboratory Medicine and
20. Rasouli N, Kern PA. Adipocytokines and the metabolic compli- Patient Care Services.
cations of obesity. J Clin Endocrinol Metab 2008;93:S64-73. Ravi Retnakaran is supported by a CIHR Clinical Re-
21. Papastefanou I, Samolis S, Panagopoulos P, et al. Correlation search Initiative New Investigator Award, Canadian Diabetes
between maternal first trimester plasma leptin levels and birth Association Clinician Scientist incentive funding and an
weight among normotensive and preeclamptic women. J Matern
Ontario Ministry of Research and Innovation Early Re-
Fetal Neonatal Med 2010;23:1435-43.
22. Mise H, Yura S, Itoh H, et al. The relationship between maternal searcher Award. Anthony Hanley holds a Tier II Canada
plasma leptin levels and fetal growth restriction. Endocr J 2007; Research Chair in Diabetes Epidemiology. Bernard Zinman
54:945-51. holds the Sam and Judy Pencer Family Chair in Diabetes
23. Verhaeghe J, van Bree R, Van Herck E. Maternal body size and Research at Mount Sinai Hospital and University of Toronto.

1360 CMAJ, September 4, 2012, 184(12)