Leptin in Cord Blood Associates with Asthma Risk at Age 3 in the Offspring of Women with Gestational Obesity

Abstract

Rationale: Maternal obesity is associated with asthma in the offspring. Whether cord blood leptin is associated with risk of asthma in offspring is unclear.

Objectives: To assess whether cord blood leptin from women with pregestational obesity predict preschool asthma.

Methods: In this birth cohort study, we divided pregnant women into three weight categories during the first obstetric visit: normal (NL), overweight (OW), and obese (OB).

Results: We followed the offspring recording atopy, wheezing, and other respiratory illnesses through 30 months of age. Cord blood and peripheral blood at 30 months of age were taken to measure cytokines, adipokines, metabolic biomarkers, and specific immunoglobulin E. Adjusted regression models were used to evaluate the association between maternal obesity and offspring asthma risk, defined by a positive Asthma Predictive Index. Three hundred thirty-nine mothers were recruited; 140 offspring were born from NL, 80 from OW, and 119 from OB mothers. OB women were older and less educated and had higher parity and higher C-section frequency. Offspring from OB women had higher birthweight, head circumference, and placental weight compared with other groups. The proportion of Asthma Predictive Index positive at 30 months of age was 12.2% in the NL, 14.7% in the OW, and 16.8% in the OB group (P = 0.18). Offspring from OB women had higher leptin, leptin/adiponectin ratio, interleukin-10, and insulin than the OW group and higher leptin than the NL group. In the adjusted analysis, offspring from OB mothers with high cord blood leptin had increased risk of asthma (adjusted odds ratio, 1.30; 95% confidence interval, 1.1–1.55; P = 0.003).

Conclusions: Offspring from obese mothers with high cord blood leptin have 30% higher asthma risk at age 3.

Clinical trial registered with ClinicalTrials.gov (NCT02903134).

Asthma is a major public health problem (1), with a global prevalence of 11.7% in children 6–7 years of age and 14.1% at 13–14 years (2), making asthma one of the most prevalent chronic diseases of childhood. In Chile, the prevalence of current wheeze among children ranges from 15.5% to 17.9% (3), and for recurrent wheezing (≥3 episodes) during the first year of life, it is as high as 22% (4).

Likewise, obesity and overweight constitute a global pandemic affecting up to 13% and 39% of the adult population, respectively (5). In Chile, the prevalence of obesity is 34.4%, including up to 38.4% in women of reproductive age (6). In a meta-analysis of 12 studies, compared with children from mothers of normal weight, those whose mothers were obese have higher odds of asthma or wheeze ever (odds ratio [OR], 1.49; 95% confidence interval [CI], 1.22–1.83; P < 0.001) and current asthma or wheeze (OR, 1.36; 95% CI, 1.08–1.68; P = 0.008) (7). Maternal overweight (body mass index [BMI] 25–30 kg/m²) showed nonsignificant trends for asthma or wheeze ever and current asthma wheeze.

Better knowledge of the in utero risk factors for asthma could offer opportunities for primary prevention (8–10). However, the mechanisms linking maternal BMI to asthma in the offspring remain poorly understood. Although a relationship between maternal BMI and childhood asthma is consistently observed (7), it is less consistent for atopic phenotypes like eczema, hay fever, or allergic sensitization (11, 12), suggesting a potential role for nonallergic pathways. In a nonselected birth cohort study in Tucson (n = 233) (13), maternal weight gain during pregnancy was associated with childhood asthma. Maternal prenatal leptin levels correlated strongly with prepregnant BMI and pregnancy weight gain, but leptin was not associated with childhood asthma.

In this study, we examine longitudinal data from the “Maternal Obesity and Asthma” birth cohort to assess whether obesity in early pregnancy is associated with asthma in preschool age. Our hypothesis is that offspring from obese mothers have an increased risk of asthma.

Methods

Population and Study Procedures

We implemented a birth cohort study (NCT02903134) recruiting pregnant women during their stay in the delivery ward at Hospital Sótero del Río in Santiago, Chile, between December 2014 and January 2016. Women were eligible if they were >18 years of age, had a singleton pregnancy, planned to seek care for their child in Santiago at least for the first 3 years, and had reliable contact information. Women with chronic conditions (diabetes, hypertension, and cardiovascular, immunologic, and neurologic diseases) were excluded. Written consent was obtained from those who agreed to participate after receiving information about the study. The study was approved by the Ethics Committee of the School of Medicine of the Pontificia Universidad Católica de Chile and by the Ethics Committee of Metropolitan Southeast Health Service of Santiago (Protocol #13-225).

Information collected at delivery included maternal age, parity, number of antenatal visits, past oral contraceptive use, paracetamol and antibiotic use during pregnancy, parental education, parental asthma and allergic diseases, number of siblings, smoking habits, home heating characteristics, carpeting, presence of pets, and dampness or mold at home. Maternal weight and height during the first antenatal visit were extracted from the obstetric chart and used to classify the cohort into three groups: normal weight (NL, BMI >18.5 kg/m² and ≤24.9 kg/m²), overweight (OW, BMI 25–30 kg/m²), or obese (OB, BMI >30 kg/m²).

After delivery and before placental discharge, 70 ml of cord blood was collected by umbilical vein puncture using cord blood collection bags (Terumo Corporation). Samples were refrigerated at 4°C until processing within 16 hours. Umbilical cord blood was centrifuged (4,500 rpm for 10 min) at room temperature; plasma was extracted from the supernatant fraction, homogenized, aliquoted, and stored at −80°C until all samples of the cohort were collected. Placentae were weighed after cutting the umbilical cord the nearest to the insertion point in the fetal side (∼2 cm) and trimming the fetal membranes. Placental efficiency was calculated dividing the neonatal weight (g) by placental weight (g) (14).

After the child’s birth, data were collected from the inpatient medical records, including sex; anthropometry; Appearance, Pulse, Grimace, Activity, and Respiration (APGAR) scores; and perinatal problems including jaundice and phototherapy. Children were followed by phone at 6, 12, 18, and 24 months of age and in person at 30 months of age. At those visits, information was collected about breastfeeding and feeding characteristics, allergic diseases (rhinitis, rhino-conjunctivitis, and atopic dermatitis), acute respiratory illnesses (bronchiolitis, acute otitis media, laryngitis or croup, and pneumonia diagnosed by chest X-ray), use of antibiotics and acetaminophen, wheezing (frequency, without cold, emergency depart consultation, hospitalization, use of albuterol, and oral/inhaled steroids), cough (at night or after crying, laughing, or agitation), siblings, and day care attendance. The Edinburg scale for maternal depression postdelivery was performed at 6 months.

Biomarkers and Cytokines

The following biomarkers and cytokines were measured in cord plasma: insulin (pg/ml) and leptin (ng/ml) were determined using the Milliplex Human Adipokine kit; interleukin (IL)-1β, IL-4, IL-6, IL-10, IL-12 p40 subunit (IL-12p40) and IL-12p70, IL-13, and tumor necrosis factor-α (all in pg/ml) were determined by Milliplex Human Cytokine/Chemokine kit; adiponectin (μg/ml) was measured by enzyme-linked immunosorbent assay (ELISA) (high molecular weight [HMW] and Total Adiponectin ELISA; Alpco); ultrasensitive C-reactive protein (mg/L), insulin (mg/dl), lipids (mg/dl) (cholesterol, triglyceride, high-density lipids, low-density lipid), and 25-hydroxyvitamin D (ng/ml) by liquid chromatography tandem mass spectrometry; and club cell secretory protein (ng/ml) by ELISA (BioVendor).

At 30 months of age, a peripheral blood sample was taken from participating children to measure white blood count, cytokines (similar to those measured in cord blood, excluding tumor necrosis factor-α and including IL-33), club cell secretory protein, adiponectin, leptin, C-reactive protein, thymic stromal lymphopoietin (pg/ml), and serum immunoglobulin E antibodies to a mix of allergens (egg, milk, wheat, nuts, soybean, fish/shellfish, dust mite, grass, weed, tree, mold/fungi, cat, and dog) with ImmunoCAP Phadiatop; a serum immunoglobulin E Phadiatop level ≥0.35 kUA/L was reported as a positive test and such children were considered atopic.

Statistical Analyses

The primary outcome was risk of asthma at age 30 months, defined as positive Asthma Predictive Index (API) according to the original stringent criteria (15): ≥3 wheezing episodes in the past year plus either ≥1 major criterion (child with diagnosis of eczema or a parental diagnosis of asthma) or ≥2 minor criteria (child with diagnosis of allergic rhinitis, wheezing apart from colds, or peripheral eosinophils ≥4%). The stringent API has been validated and is widely used in epidemiological studies of asthma (16).

Biomarkers (cytokines, adipokines, and metabolic markers) were normalized using log (blood level +0.25); undetectable values were assigned 1/4 of the lower limit of detection for each assay. Results were presented as geometric mean taking exponential of the mean of log values. Bivariate analyses were performed using Fisher’s exact test for categorical variables and t tests or analysis of variance for continuous variables. If Levene’s test for equality of variances was nonsignificant, we considered analysis of variance to be adequate, independent of the normality of the original variables, given the sample size in each group; post hoc pairwise comparisons were performed using the Hochberg method. If variance heterogeneity was significant, we used Welch’s unequal variance test and the post hoc pairwise comparisons were performed via the Games-Howell approach.

Multivariable analyses to evaluate associations between maternal obesity groups and risk of asthma were performed using logistic regression models; parameters and confidence intervals were estimated by maximum likelihood. Potential confounders like sex, type of delivery, placental weight, birthweight, jaundice, maternal age, parental asthma, parental education, smoking at home, pets, siblings, and type of home heating were included in the model and were retained in the final models if their coefficients were significant or they contributed significantly to the best model fit. In addition, we evaluated effect modification by maternal obesity using interaction terms (e.g., obesity × cytokine). Leptin was categorized in quartiles and the leptin’s 25th percentile was the reference group. A P value <0.05 was considered statistically significant. SPSS v17.0 (IBM) was used throughout.

Sample size for the cohort was calculated based in the main outcome of the study: asthma risk (i.e., positive API). Asthma incidence in Chile is estimated at 15% (3), and based on previous studies, we predicted 30% higher odds of asthma in the offspring of obese women (7). Assuming a two-tailed test, 80% power, and a 5% alpha, we needed 120 women per group (OB and NL).

Results

A total of 339 pregnant women were enrolled in the Santiago Cohort during 2014–2016. Among these, 140 offspring were born from NL mothers, 80 from OW mothers, and 119 from OB mothers; mean BMIs were 22.7 kg/m², 27.3 kg/m², and 34.2 kg/m², respectively. OB women were significantly older and less educated and had higher parity, and a higher proportion of them underwent C-section (Table 1). Fathers from the OB group were also less educated. There were no differences in the number of antenatal visits, gestation week at the first and last visit, past oral conceptive use, or use of paracetamol or antibiotics during pregnancy (Table 1). Likewise, home characteristics were similar between groups, as was the Edinburg total score for postdelivery maternal depression (Table 1).

Table 1.

Parental, home, and neonatal baseline characteristics by maternal BMI group

Characteristics	Normal Weight	Overweightn = 80	Obesen = 119	P Value*
	n = 140
Maternal data
Age, yr	23.55 ± 5.3	25.20 ± 5.5	28.19 ± 5.8	<0.001
Parity	1 [1–2]	2 [1–3]	2 [2–3]	<0.001
C-section	17.1%	26.3%	37.8%	0.003
Weight at first antenatal visit, kg	57.90 ± 5.7	68.86 ± 7.4	85.45 ± 11.8	<0.001
Weight at last antenatal visit, kg	70.10 ± 7.5	79.50 ± 8.2	94.30 ± 12.6	<0.001
BMI, kg/m2	22.68 ± 1.6	27.30 ± 1.8	34.17 ± 3.6	<0.001
Number of antenatal visits	6.91 ± 1.9	7.11 ± 1.6	6.70 ± 1.9	0.29
Paracetamol during gestation	18.8%	27.8%	29.1%	0.12
Antibiotics during gestation	25.9%	29.1%	27.1%	0.88
Highest education level				0.01
Basic	17.3%	17.5%	23.7%
Secondary	62.6%	62.5%	66.9%
Higher	20.1%	20.0%	9.3%
Maternal asthma	11.5%	16.7%	14.5	0.52
Maternal rhinitis	17.3%	24.1%	22.0	0.42
Maternal dermatitis	7.2%	10.3%	8.5	0.72
Edinburgh depression scale	7.42 [6.3–8.6]	8.13 [6.6–9.7]	7.27 [6.2–8.3]	0.63
Paternal data
Highest education level				0.009
Basic	16.3%	11.7%	19.5%
Secondary	54.8%	68.8%	69.0%
Higher	28.9%	19.5%	11.5%
Asthma	11.4%	12.2%	8.9%	0.75
Rhinitis	21.8%	17.3%	24.6%	0.51
Dermatitis	10.9%	10.7%	8.0%	0.72
Home data
Pets inside house	36.4%	31.3%	27.7%	0.32
Smoking inside house	12.2%	7.6%	12.7%	0.52
Smoking outside	63.3%	57.0%	59.3%	0.63
Home heating				0.37
None	12.2%	12.7%	11.0%
Electric	21.6%	13.9%	24.6%
Gas	33.1%	38.0%	32.2%
Chiminea	12.9%	6.3%	6.8%
Kerosene	20.1%	29.1%	25.4%
Dampness and mold	22.1%	27.5%	21.8%	0.60
Use of sprays for cleaning	55.7%	66.3%	54.2%	0.59
Sibling or other children at home	63.2%	70.6%	70.6%	0.45
Neonatal data
Gestational age, wk	38.94 ± 1.1	39.02 ± 1.2	38.92 ± 1.1	0.79
Sex, M	52.9%	53.8%	51.3%	0.94
Weight at birth, g	3,424.3 ± 404	3,445.4 ± 424	3,573.1 ± 512	0.022
Height at birth, cm	49.67 ± 1.8	49.59 ± 1.9	49.56 ± 2.3	0.90
Head circumference, cm	34.43 ± 1.6	34.70 ± 1.4	35.06 ± 1.2	0.002
Respiratory problems	1.4%	1.3%	1.7%	1.0
Jaundice	15.0%	18.3%	21.2%	0.55
Phototherapy	4.4%	7.0%	7.7%	0.58
Placenta weight, g	417.4 ± 82.2	408.6 ± 88.3	445.4 ± 90.1	0.007

Definition of abbreviation: BMI = body mass index.

Numbers expressed as mean ± standard deviation, median [25–75 percentile], or percent (categorical variables). Bold typeface indicates statistical significance (P < 0.05). Total N = 339.

^*P value for trend test analysis of variance (continuous variables) or Fisher’s exact test (categorical variables).

Offspring from the OB group had higher birthweight, head circumference, and higher proportion of placental weight compared with the offspring of other groups (Table 1). There were no differences in newborn sex distribution; gestational age; birth length; Appearance, Pulse, Grimace, Activity, and Respiration score; neonatal problems; or jaundice and phototherapy.

Follow-Up Data at 6 through 30 Months of Age

Upon follow-up at 6 months, the age at which complementary foods were started was earlier in the OB group. At 12 months, infants from the OB group were more likely to live in homes with other children (Table 2). There were no differences in atopic diseases (rhinitis, rhinoconjuntivitis, and dermatitis) between groups in the follow-up at ages 6–30 months, with the only exception of higher prevalence of rhinitis in the offspring from the OW group at 18 months and higher dermatitis in the OB group at 24 months (Table 2). The accumulated prevalence of bronchiolitis was significant higher in the OB group at 30 months (Table 2). The overall proportion of positive API at 30 months of age was 14.5%: 12.2% in the NL group, 14.7% in the OW group, and 16.8% in the OB group, although the differences were not significant (P = 0.62). The proportions of different API major and minor criteria were similar among groups (see Table E1 in the online supplement). Similarly, there were no differences between groups in respiratory illness (acute otitis media, adenoiditis, croup, pneumonia by chest X-ray), use of paracetamol, or use of antibiotics for respiratory and for nonrespiratory illnesses at 6–30 months of age, with the exception of more frequent paracetamol use at 12 months among the offspring from the OW group (see Table E2).

Table 2.

Offspring characteristics at 6–30 months of age, by maternal BMI group

Characteristics	Normal Weight	Overweightn = 68	Obesen = 102	P Value*
	n = 114
Survey at 6 mo (n = 284)
Age at questionnaire, mo	6.46 ± 0.7	6.37 ± 0.6	6.48 ± 0.8	0.60
Age at start of day care, yr	0.40 ± 1.5	0.33 ± 1.3	0.13 ± 0.9	0.22
Exclusive breastfeeding, mo	4.45 ± 2.2	4.31 ± 2.3	4.03 ± 2.3	0.41
Age start complementary food, mo	5.30 ± 1.7	4.89 ± 2.1	5.60 ± 1.2	0.02
Accumulated prevalence ofbronchiolitis	34.7%	40.0%	31.5%	0.49
Any wheezing	34.5%	33.8%	30.4%	0.82
Hospitalization for wheezingepisode	9.8%	7.4%	8.8%	0.87
ER for wheezing episodes	23.2%	19.1%	22.5%	0.84
Other children at home	63.2%	70.6%	70.6%	0.45
Allergic rhinitis	22.1%	25.0%	30.4%	0.40
Allergic rhinoconjunctivitis	8.8%	13.4%	18.6%	0.11
Atopic dermatitis	7.1%	8.8%	13.7%	0.27
Survey at 12 mo (n = 288)	n = 119	n = 68	n = 101
Age at questionnaire	12.48 ± 0.8	12.41 ± 0.9	12.39 ± 0.8	0.68
Accumulated prevalence of bronchiolitis	58.7%	65.8%	58.5%	0.56
Any wheezing	46.2%	51.5%	42.6%	0.54
Hospitalization for asthma/wheezing	9.3%	8.8%	5.0%	0.43
ER for asthma/wheezing	33.1%	33.8%	31.7%	0.96
Other children at home	57.1%	67.6%	74.3%	0.027
Allergic rhinitis	29.4%	38.2%	29.7%	0.42
Allergic rhinoconjunctivitis	17.6%	27.9%	21.8%	0.26
Atopic dermatitis	5.9%	7.4%	8.9%	0.68
Survey at 18 mo (n = 284)	n = 117	n = 67	n = 100
Age at questionnaire, mo	18.71 ± 1.1	18.40 ± 1.1	18.72 ± 1.1	0.121
Accumulated prevalence of bronchiolitis	65.5%	74.0%	68.0%	0.47
Any wheezing	37.1%	50.7%	45.0%	0.175
Hospitalization for asthma/wheezing	3.5%	4.8%	5.0%	0.806
ER for asthma/wheezing	27.2%	27.0%	37.0%	0.239
Other children at home	67.2%	78.5%	73.0%	0.264
Allergic rhinitis	31.0%	44.8%	25.3%	0.032
Allergic rhinoconjunctivitis	19.1%	24.2%	14.1%	0.26
Atopic dermatitis	10.3%	14.9%	12.0%	0.62
Survey at 24 mo (n = 258)	n = 104	n = 63	n = 91
Age at questionnaire, mo	24.41 ± 0.9	24.78 ± 1.2	24.63 ± 1.0	0.077
Accumulated prevalence of bronchiolitis	69.0%	82.9%	74.0%	0.11
Any wheezing	51.9%	66.7%	56.0%	0.167
Hospitalization for asthma/wheezing	2.9%	3.2%	0.0%	0.216
ER for asthma/wheezing	18.3%	19.0%	20.2%	0.979
Allergic rhinitis	26.9%	38.7%	31.9%	0.29
Allergic rhinoconjunctivitis	12.5%	25.8%	14.4%	0.075
Atopic dermatitis	18.3%	29.0%	35.2%	0.025
Survey at 30 mo (total n = 233)	n = 103	n = 51	n = 79
Age at questionnaire, mo	30.20 ± 1.1	30.04 ± 1.0	30.13 ± 1.0	0.634
Wheezing/asthma	50.5%	45.1%	59.5%	0.243
Accumulated prevalence of bronchiolitis	70.5%	85.3%	81.9%	0.04
Hospitalization for asthma/wheezing	5.8%	0.0%	1.3%	0.138
ER for asthma/wheezing	11.7%	13.7%	21.5%	0.190
Allergic rhinitis	27.2%	23.5%	35.8%	0.297
Allergic rhinoconjunctivitis	17.5%	15.7%	21.5%	0.707
Atopic dermatitis	15.7%	21.6%	27.8%	0.143
Positive API	12.2%	14.5%	16.8%	0.184

Definition of abbreviations: API = Asthma Predictive Index; BMI = body mass index; ER = emergency room.

Numbers expressed as mean ± standard deviation or as a percent (categorical variables). Bold typeface indicates statistical significance (P < 0.05).

^*P value for trend test analysis of variance (continuous variables) or Fisher’s exact test (categorical variables).

Biomarkers in Cord Blood and in Peripheral Blood at Age 30 Months

Cord blood samples were available for 324 of 339 (95.6%) children. Leptin levels, leptin/adiponectin ratio, IL-10, and insulin were each significantly different among maternal BMI groups. Children from OB mothers had higher leptin than children from the OW and NL groups and higher leptin/adiponectin ratio, IL-10, and insulin than those from OW mothers (Table 3). There were no significant differences among groups for the rest of the biomarkers in cord blood. At the 30-month visit (n = 164), there were no significant differences in biomarker levels among the three maternal BMI groups (see Table E3), and the prevalence of atopy (positive detectable Phadiatop and % peripheral eosinophils) was also similar (Phadiatop: 21.3% in NL vs. 33.3% in OW vs. 26.4% in OB women, P = 0.64; and % eosinophils: 2.74 ± 1.9 vs. 2.93 ± 2 vs. 3.08 ± 1.98, P = 0.626, respectively).

Table 3.

Biomarkers from cord by maternal BMI group

Cord Blood	Normal Weight	Overweight	Obese	P Value*
	n = 133	n = 78	n = 113
Insulin, mg/dl	46.24 ± 3.54	37.12 ± 4.17†	58.64 ± 3.05†	0.046
Adiponectin, μg/ml	9.24 ± 1.5	10.04 ± 1.51	10 ± 1.51	0.222
Leptin, ng/ml	5.67 ± 3.04†	5.04 ± 3.3†	8.61 ± 3.29†‡	0.003
Leptin/adiponectin	0.96 ± 2.33	0.84 ± 2.5†	1.22 ± 2.44†	0.011
Club cell 16, ng/ml	4.01 ± 1.97	3.45 ± 2.01	3.64 ± 2.37	0.332§
IL-β1, pg/ml	2.22 ± 2.71	1.8 ± 2.91	2.56 ± 2.83	0.069
IL-4, pg/ml	8.95 ± 2.51	8.75 ± 2.23	11.01 ± 2.14	0.092
IL-6, pg/ml	6.36 ± 3.45	6.31 ± 3.54	6.99 ± 3.5	0.8
IL-10, pg/ml	1.74 ± 3.13	1.44 ± 2.81†	2.29 ± 2.49†	0.005
IL-12 p40, pg/ml	26.67 ± 2.25	25.45 ± 1.56	30.58 ± 1.92	0.143
IL-12 p70, pg/ml	8.16 ± 2.8	8.09 ± 2.46	9.52 ± 2.43	0.373
IL-13, pg/ml	11.12 ± 2.43	11.09 ± 1.82	10.92 ± 2.71	0.986
TNF-alfa, pg/ml	7.34 ± 1.93	7.22 ± 1.59	8.12 ± 1.51	0.231
usCRP	UD	UD	UD
Cholesterol, mg/dl	36.22 ± 1.55	32.09 ± 1.66	34.93 ± 1.61	0.196
Triglycerides, mg/dl	26.63 ± 1.91	24.89 ± 2.03	25.8 ± 1.86	0.767
HDL, mg/dl	7.25 ± 2.07	6.69 ± 2.14	6.35 ± 2.22	0.39
LDL, mg/dl	20.37 ± 1.92	17.38 ± 2.13	20.62 ± 1.91	0.176
Vitamin D total	19.95 ± 1.46	19.71 ± 1.58	21.93 ± 1.45	0.094§

Definition of abbreviations: BMI = body mass index; HDL = high-density lipids; IL = interleukin; LDL = low-density lipids; TNF = tumor necrosis factor; UD = undetectable; usCRP = ultrasensitive C-reactive protein.

Numbers expressed as geometric mean ± standard deviation or as a percent (categorical variables). All biomarkers and cytokines transformed as log(level+0.25) to normalize distribution. Log-transformed values used for all analyses. For ease of interpretation, tables show the geometric means and standard deviation. Bold typeface indicates statistical significance (P < 0.05). Total N = 324.

^*P value for trend by analysis of variance test.

^†Significantly different in overweight versus obese group.

^‡Significantly different than normal-weight group.

^§P value for robust Welch test (unequal variances).

Adjusted Analyses of Maternal Weight, Leptin, and Asthma Risk

We performed multivariate logistic regression models using the forward method for variable selection with asthma risk (positive API) at 30 months of age as the dependent variable. Our main independent variables included maternal weight group and biomarkers of obesity (leptin, leptin/adiponectin, IL-10, insulin in cord blood). After adjusting for maternal education and smoking at home, higher cord blood leptin levels in the offspring from OB mothers was associated with increased risk of asthma: the odds of positive API increased by 30% for each 10 ng/ml higher leptin (adjusted OR, 1.30; 95% CI, 1.1–1.55; P = 0.003), (Table 4). The 25th percentile of leptin (“low” leptin) was 2.44 ng/ml, whereas the 75th percentile (“high” leptin) was 13.64 ng/ml. Compared with children with low leptin, those with high leptin had a 34.5% higher risk of asthma (adjusted OR, 1.345; 95% CI, 1.11–1.63; P = 0.003).

Table 4.

Logistic regression model for positive API at 30 months of age

	Beta	SE	OR	95% CI	P Value
Constant	−3.205	0.614	0.041	—	<0.001
Leptin in NW group			Reference
Leptin in OW group*	−0.32	0.39	0.73	0.34–1.56	0.41
Leptin in OB group*	0.265	0.088	1.30	1.1 to 1.55	0.003
Tobacco inside house	0.949	0.543	2.55	0.89–7.50	0.081
Maternal education primary			Reference
Maternal education secondary	1.223	0.634	3.40	0.98–11.77	0.054
Maternal education university	1.584	0.711	4.87	1.21–19.62	0.026

Definition of abbreviations: API = Asthma Predictive Index; CI = confidence interval; NW = normal weight; OB = obese; OR = odds ratio; OW = overweight; SE = standard error.

Bold typeface indicates statistical significance (P < 0.05). Total N = 339.

^*Shown are the coefficient, adjusted OR, and 95% CI for the association between leptin and positive API in each group of maternal body mass index.

Discussion

In this birth cohort, we describe, for first time, that offspring from obese mothers with high levels of leptin in cord blood have higher risk of asthma at 2.5 years of age. Indeed, among children born to obese mothers, the risk of having a positive API increased by ∼30% for each 10 ng/ml higher leptin. Because maternal obesity is a modifiable risk factor, our results add to the growing evidence that a public health effort to prevent obesity in women, particularly those of reproductive age, should be emphasized.

Although the precise mechanisms are not fully understood, several epidemiological studies have linked maternal obesity to childhood asthma. A meta-analysis of 12 studies by Forno and colleagues (7) reported an association between maternal obesity and childhood asthma up to 16 years of life, independent of sex, with nonsignificant trends for maternal overweight. More recently, a longitudinal study on maternal obesity and offspring asthma in U.S. children (17) reported that maternal prepregnancy overweight (adjusted OR, 1.19; 95% CI, 1.03–1.38) and obesity (adjusted OR, 1.34; 95% CI, 1.08–1.68) were associated with offspring asthma up to about the age of 12 years; the association was observed with nonallergic asthma only in boys, whereas allergic asthma was seen only in girls. Moreover, a recent birth cohort in Finland (18) reported that male offspring of mothers with higher BMI had more asthma in adult life (adjusted hazard ratio, 1.54; 95% CI, 1.18–2.02) compared with those born from mothers with lower BMI; however, that association was not present in women.

Maternal obesity may affect long-term offspring health via several pathways, including excess nutrients, elevated levels of hormones, and elevated inflammatory mediators (19). Our study shows that the effect on the risk of childhood asthma is present since early life and may be mediated by leptin. High prepregnancy BMI, along with excessive gestational weight gain in the second and third trimesters, have been associated with higher levels of cord blood leptin (20, 21). Leptin is an adipokine (a type of cytokine secreted from adipose tissue) with systemic proinflammatory effects, including increased interferon-γ–mediated responses, increased CD4⁺ T-cell immunity and activation of mast cells, as well as activation of transcription factors such as nuclear factor-κB. It may also have proinflammatory effects on the airways (13, 22). Leptin concentration positively correlates with BMI in pregnant women, with ponderal index in newborns, and with placental weight (23). However, leptin levels in nonobese mothers are not affected by the nutritional status of the fetus. Large-for-gestational-age neonates born from mothers with BMI >25 kg/m² had higher leptin than large-for-gestational-age neonates born from mothers with normal BMI, whereas in adequate-for-gestational age neonates, leptin is similar regardless of the nutritional status of their mothers—suggesting that leptin does not cross the placenta (23). In our cohort, higher leptin in cord blood was associated with asthma risk only among the offspring of obese mothers; this suggests that although leptin may be an important factor or biomarker for asthma risk in these children, other elements of maternal obesity may be important as well. We found significantly higher insulin levels in the cord blood from OB mothers. Prior studies have shown that maternal hyperglycemia is associated with higher cord blood insulin (24) and that maternal glycemia during pregnancy is associated with hypomethylation of the leptin gene (LEP) in neonates, which in turn leads to higher leptin levels in cord blood (25).

Our study has some limitations. First, given the sample size of the cohort, we likely lacked power to detect small effects. Because of the cohort’s age, we used the API rather than a physician diagnosis of asthma. However, the stringent API has been extensively used and validated in epidemiological studies. Anthropometric data of children at 30 months of age were available only for a small subset. At the same time, the study has several strengths. Data and samples were collected and processed prospectively, following standardized protocols. It includes very detailed baseline characteristics and phenotyping every 6 months through age 30 months, a crucial time for the maturation of the immune system and the development of atopic diseases.

If confirmed, our findings could have significant implications. Maternal obesity is a risk factor for multiple pregnancy complications that affect both mother and child. As a modifiable risk factor, prepregnancy obesity should be targeted in preconception programs that promote a healthy weight before and throughout pregnancy. Although no single risk factor can entirely account for childhood asthma, such a prevention strategy may reduce the incidence of early childhood asthma in future generations (26). More importantly, leptin levels may constitute an important predictive biomarker for asthma risk among offspring of obese women and thus would allow us to identify women at higher risk, in whom interventions may be more crucial.

Conclusions

Offspring born from obese mothers with high cord blood levels of leptin have higher risk of asthma at 30 months of age. Public health efforts trying to prevent obesity in women of reproductive age should be emphasized and may lead to reduced asthma risk in their offspring.