Maternal Metabolic Biomarkers are Associated with Obesity and Excess Gestational Weight Gain

Abstract

Objective

The purpose of this study was to evaluate the independent contribution of maternal obesity and gestational weight gain (GWG) in excess of the Institute of Medicine’s guidelines on levels of maternal serum inflammatory and metabolic measures.

Study Design

Banked maternal serum samples from 120 subjects with documented prepregnancy or first trimester body mass index (BMI) were utilized for analyte analyses. Validated, BMI-specific formulas were utilized to categorize GWG as either insufficient, at goal or excess based on the Institute of Medicine guidelines with gestational age adjustments. Serum was analyzed for known inflammatory or metabolic pathway intermediates using the Luminex xMap system with the MILLIPLEX Human Metabolic Hormone Magnetic Bead Panel. Measured analytes included interleukin-6, monocyte chemoattractant protein-1, and tumor necrosis factor-α and metabolic markers amylin, c-peptide, ghrelin, gastric inhibitory polypeptide, glucagon-like peptide-1, glucagon, insulin, leptin, pancreatic polypeptide, and peptide YY. Kruskal–Wallis ANOVA and Pearson’s correlation coefficients were calculated for each marker.

Results

C-peptide, insulin, and leptin all varied significantly with both obesity and GWG while glucagon-like peptide −1 varied by BMI but not GWG. These analytes covaried with other metabolic analytes, but not with inflammatory analytes.

Conclusion

Maternal metabolic biomarkers at delivery vary significantly with both obesity and GWG. Taken together, these findings suggest that GWG (with and without comorbid obesity) is an important mediator of measurable metabolites in pregnancy but is not necessarily accompanied by inflammatory measures in serum. These findings are consistent with GWG being an independent risk factor for metabolic disturbances during pregnancy.

Maternal obesity is associated with and causal to a myriad of adverse pregnancy and neonatal outcomes and can have lasting impact on childhood obesity and (later) adult obesity.¹ Studies have also shown adverse pregnancy outcomes associated with excessive gestational weight gain (GWG), including hypertensive disorders of pregnancy and preterm delivery, both indicated and spontaneous.¹ While the etiology of these conditions are multifactorial, their association with inflammation is well established.^2,3

Inflammatory and metabolic markers have been shown to be altered among nonpregnant individuals with obesity; specifically, a pro-inflammatory, insulin resistant state is observed.^4-8 While variations in placental^9,10 and systemic^11-13 inflammation have been observed previously among gravidae with obesity, assessment by GWG (in both the presence and absence of obesity) until recently has been limited.¹⁴

In this study, we sought to evaluate the independent contribution of maternal body mass index (BMI) and GWG on maternal serum levels of inflammatory and metabolic markers measured at the time of term delivery. Our hypothesis was that obesity and gestational weight gain would result in aberrations of different inflammatory or metabolic markers as responses to different exposures (established adipose tissue vs. subacute nutritional intake).

Materials and Methods

This study was approved by the institutional review board (IRB) for Baylor College of Medicine (institutional review board H-26589). All subjects in this study were enrolled into Baylor College of Medicine’s Universal Perinatal Database and Biospecimen Repository (PeriBank) (IRB H-26364).¹⁵ Complete details of the protocol for the PeriBank database and biospecimen repository are described in detail elsewhere.¹⁵ Briefly, all gravidae who present to Ben Taub Hospital and Texas Children’s Hospital Pavilion for Women for anticipated delivery are approached regarding enrollment into PeriBank; 85 to 93% of gravidae approached consent to participate.¹⁵ For those who provide informed consent, data are extracted from the electronic medical record and additional data points are obtained via questionnaires. Biospecimens, such as maternal blood, paternal blood, umbilical cord blood, and placental samples, are collected.¹⁵

For the purpose of this study, samples from 120 PeriBank participants were obtained with 40 participants from each of the three classifications of BMI described below. Extracted and analyzed clinical metadata included self-reported and/or documented prepregnancy weight (as recorded in the electronic medical record), documented early pregnancy weight (defined as first trimester weight), documented maternal delivery weight, and documented maternal height. Subjects lacking any of these data (thus precluding the ability to calculate prepregnancy or early pregnancy BMI and GWG) were excluded. Baseline BMI was determined with the use of the prepregnancy or first trimester weight. In this dataset prepregnancy weight was available for 61 subjects; first trimester weight was used for the remaining 59.

All subjects were then categorized by BMI categories and GWG categories in two separate manners; thereby, the same cohort was categorized in two separate manners. First, to determine whether degree of prepregnancy overweight or obesity contributed to these analytes, samples were grouped into three BMI categories: normal weight (BMI: 18.5–24.9 kg/m²), overweight (BMI: 25–29.9 kg/m²), and obese (BMI ≥ 30 kg/m², all classes). GWG was calculated by subtracting the prepregnancy or early pregnancy weight from the weight documented at the time of delivery. To calculate whether GWG exceeded recommendations, we employed previously validated and BMI-specific formulas derived from the Institute of Medicine (IOM).^16-18 Excess GWG was defined as weight gain that was in excess of the maximum recommended weight gain for a given gestational age and prepregnancy BMI. Insufficient GWG was defined as GWG below the lowest recommended GWG per IOM following the formulas mentioned above.^16-18 These categorizations were used to classify GWG as “insufficient GWG,” sufficient, or “goal GWG,” and “excess GWG.”

We also secondarily sought to assess gravidae with or without obesity and also those with or without excess GWG. Thus, the same samples were reclassified into two such groups with designations of nonobese (BMI < 30 kg/m²) and obese (BMI ≥ 30 kg/m²), and also those with excess GWG (“excess GWG”) and those without (“no excess GWG”). For the purpose of this second analysis, subjects whose GWG was less than recommended were grouped with those without excess GWG. Finally, we assessed the correlation of BMI and GWG as continuous variables with each analyte.

Because this study investigated inflammatory and metabolic markers, to minimize confounding, subjects with diabetes (prepregnancy or gestational), asthma, chronic hypertension, any hypertensive disorder of pregnancy, systemic lupus erythematous, multifetal gestation, underweight (BMI < 18.5 kg/m²), intraamniotic infection, and endometritis were excluded.

All serum samples were collected under uniform protocol by personnel trained on our universal perinatal database and biospecimen repository protocol, strictly following PeriBank’s detailed manual of procedures.¹⁵ All clinical metadata were extracted from the electronic medical record or recorded via patient interview as described in PeriBank’s protocol.¹⁵

For this study, the database was first queried for participants with serum samples available who met inclusion criteria. Once 120 specimens were randomly selected from the subjects who met inclusion criteria for the study, the specimens were obtained from PeriBank’s −80°C storage freezers that is processed and analyzed using Luminex xMap Milliplex multiplex.

Luminex xMap Technology

In this study, the EMD Millipore MILLIPLEX Human Metabolic Hormone Magnetic Bead Panel was used to measure interleukin-6(IL-6), tumor necrosis factor-α (TNF-α), monocyte chemoattractant protein-1 (MCP-1), c-peptide, ghrelin, insulin, leptin, amylin, gastric inhibitory polypeptide (GIP), glucagon-like peptide (GLP-1), glucagon, pancreatic polypeptide (PP), and peptide YY (PYY) and the assays were read on Luminex 100 (Millipor-eSigma, Merck, Darmstadt, Germany). The MILLIPLEX Human Metabolic Hormone Magnetic Bead Panel is a 15 analytes, 13-plex kit used for the simultaneous quantification of any or all 13 aforementioned analytes in serum, plasma, and cell/tissue culture samples. Each maternal serum aliquot was run in duplicates according to the manufacturer’s instructions.¹⁹ There were a total of three plates. One plate was run the first day, and concentrations were calculated from the included standards. Plates 2 and 3 were run together on the subsequent day, and used standards located on each plate (2 + 3) combined to generate the curves. Because samples were run on two separate days, two curves and therefore two different sets of values for the lower limit of detection were utilized. The duplicate concentrations were averaged. For the purpose of this analysis, values that were within the limits of detection of the curve or extrapolated were utilized. Extrapolated values were defined as values below the lowest value on the curve but above the lowest limit of detection. Any subject with quantification for any marker that was outside the limits of detection or out of range was excluded in the data analysis portion.

Statistical Methods

Statistical analysis was performed using SPSS 25.0 (IBM Corp; Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY). Pearson’s Chi-square test or Fisher’s exact test was performed on clinical categorical variables, and t-test and ANOVA were performed on continuous variables where appropriate. Kruskal–Wallis ANOVA, Mann–Whitney U test, and Pearson’s correlation were used to analyze Luminex analyte results by BMI and GWG. A p-value <0.05 was considered statistically significant. Because of the correlation between the different biomarkers, there was no adjustment for multiple comparisons.

Results

Samples from 120 gravidae were included in this analysis. After obtaining the samples, clinical validation resulted in reclassification of the BMI of two gravidae: one gravida originally classified as overweight was found to have an earlier weight confirming normal prepregnancy BMI and one gravida originally classified as having obesity was found to have an earlier weight confirming overweight BMI. Thus, 41 samples were obtained from gravidae with normal BMI, 40 from overweight gravidae, and 39 from gravidae with obesity. In sum, five participants had insufficient GWG, 41 had goal GWG, and 74 had excess GWG (Fig. 1).

Fig.1.

Of 120 samples, 41 samples were obtained from gravidae with normal BMI, 40 from overweight gravidae, and 39 from gravidae with obesity. In sum, five participants had insufficient GWG, 41 had goal GWG, and 74 had excess GWG. BMI, body mass index; GWG, gestational weight gain.

BMI varied with age, gravidity, and parity (p < 0.001, p < 0.001, and p < 0.004, respectively). Excess GWG was more common among gravidae of Black or Asian race (p = 0.039; Table 1). When reanalyzed as only two groups for BMI and GWG, the maternal age, race, ethnicity, parity, gestational age at delivery, and infant birthweight did not significantly differ by BMI or GWG group. (Supplementary Table 1 [available in the online version]). However, gravidae with obesity had higher gravidity (p = 0.010).

Table 1.

Maternal demographic characteristics

	Body mass index n = 120				Gestational weight gain n = 120
	Normal n = 41	Overweight n = 40	Obese n = 39	p-Value	Insufficient n = 5	Normal n = 41	Excessive n = 74	p-Value
Age (mean, SD)	24.4 (5.2)	30.6 (6.8)	29.7 (5.4)	<0.001	25.6 (8.8)	29.2 (6.2)	27.8 (6.3)	0.34
Race (n %)
White	33 (80.5)	35 (87.5)	35 (89.7)	0.19	4 (80.0)	66 (89.2)	33 (80.5)	0.039
Black or African American	2 (4.9)	3 (7.5)	4 (10.3)		0 (0)	5 (6.8)	4 (9.8)
Asian	4 (9.8)	2 (5.0)	0 (0)		0 (0)	3 (4.1)	3 (7.3)
Not reported	2 (4.9)	0 (0)	0 (0)		1 (20.0)	0 (0)	1 (2.4)
Ethnicity
Hispanic	35 (85.4)	34 (85.0)	34 (87.2)	0.96	5 (100)	64 (86.5)	34 (82.9)	0.57
Non-Hispanic	6 (14.6)	6 (15.0)	5 (12.8)		0 (0)	10 (13.5)	7 (17.1)
Gravida (n %)
1	19 (46.3)	8 (20.0)	3 (7.7)	<0.001	2 (40.0)	8 (19.5)	20 (27.0)	0.76
2	13 (31.7)	6 (15.0)	12 (30.8)		1 (20.0)	10 (24.4)	20 (27.0)
3 or more	9 (22.0)	26 (65.0)	24 (61.5)		2 (40.0)	23 (56.1)	34 (45.9)
Parity
0	21 (51.2)	11 (27.5)	8 (20.5)	0.004	2 (40.0)	13 (31.7)	25 (33.8)	0.65
1	13 (31.7)	8 (20.0)	13 (33.3)		2 (40.0)	9 (22.0)	23 (31.1)
2 or more	7 (17.1)	21 (52.5)	18 (46.2)		1 (20.0)	19 (46.3)	26 (20.0)
Gestational age at delivery (wk) (mean, SD)	39.7 (1.3)	39.3 (1.1)	39.5 (1.5)	0.48	38.9 (2.0)	39.6 (1.2)	39.5 (1.3)	0.62
Birthweight (g) (mean, SD)	3,278 (486)	3,416 (499)	3,402 (628)	0.45	2,925 (1046)	3,359 (465)	3,397 (530)	0.17

Abbreviation: SD, standard deviation.

Note: Bold indicates statistical significance.

The MILLIPLEX Human Metabolic Hormone Magnetic Bead Panel yielded results that were within detectable range or extrapolated (within the limits of detection as defined above) for all samples for amylin, c-peptide, ghrelin, GIP, glucagon, insulin, leptin, and MCP-1. Results were out of measurable range for GLP-1 for 75 samples, IL-6 for 41 samples, PP for 24 samples, and PYY for 66 samples. Assessment of inflammatory and metabolic markers by BMI (organized into three groups) demonstrated that c-peptide, insulin, and leptin increased with BMI categories (Table 2). C-peptide and insulin varied with GWG but were generally lower among gravidae with goal GWG. Leptin levels increased across GWG groups (Table 2).

Table 2.

Cytokines and metabolic analytes by body mass index and gestational weight gain groups (Kruskal–Wallis’s test)

	Body mass index				Gestational weight gain
	BMI < 25	BMI 25–29.9	BMI ≥ 30	p-Value	Insufficient GWG	Goal GWG	Excess GWG	p-Value
Analyte	n = 41	n = 40	n = 39		n = 5	n = 41	n = 74
Cytokines
IL-6	56.9 (117)	103 (277)	27.5 (58.4)	0.082	48.3 (58.5)	43.2 (63.9)	74.5 (222)	0.797
MCP-1	299 (416)	363 (341)	368 (536)	0.214	238 (137)	291 (294)	378 (506)	0.452
TNF-α	4.75 (3.88)	5.14 (3.39)	11.2 (35.4)	0.254	4.55 (1.28)	4.43 (1.90)	8.55 (26.0)	0.567
Metabolic analytes
Amylin	50.6 (31.0)	59.6 (48.3)	57.3 (51.5)	0.448	52.0 (26.5)	53.0 (52.7)	57.6 (40.2)	0.319
c-peptide	2,848 (2,118)	4,007 (2,314)	3,924 (2,806)	0.029	3,554 (2,083)	2,888 (2,292)	3,971 (2,517)	0.037
Ghrelin	25.6 (1.79)	25.6 (1.79)	26.0 (1.57)	0.276	26.0 (1.61)	25.5 (1.85)	25.8 (1.65)	0.640
GIP	57.7 (64.9)	57.1 (55.9)	56.4 (60.7)	0.906	28.6 (29.6)	50.1 (57.0)	62.8 (62.9)	0.200
GLP-1a	10.2 (26.4)	0 (0)	8.3 (23.0)	0.019	0 (0)	1.80 (8.29)	8.99 (25.1)	0.172
Glucagon	28.9 (10.8)	27.7 (6.6)	28.0 (6.4)	0.976	26.9 (3.39)	26.9 (6.5)	29.0 (9.14)	0.436
Insulin	879 (804)	1,593 (1,696)	1,644 (2,333)	0.023	1521 (1,826)	1,157 (1,896)	1,471 (1,655)	0.024
Leptin	20,893 (11,526)	29,697 (18,342)	35,789 (26,384)	0.005	17,126 (8,577)	22,295 (15,192)	32,980 (22,254)	0.001
PP	70.5 (83.8)	77.2 (90.2)	63.5 (71.6)	0.900	27.7 (45.7)	71.4 (79.8)	72.8 (84.8)	0.323
PYY	43.8 (77.6)	38.2 (81.0)	48.7 (91.9)	0.403	10.9 (24.3)	55.5 (98.4)	39.1 (75.6)	0.613

Abbreviations: BMI, body mass index; GIP, gastric inhibitory polypeptide; GLP-1, glucagon-like peptide; GWG, gestational weight gain; IL-6, interleukin-6; MCP-1, monocyte chemoattractant protein-1; PP, pancreatic polypeptide; PYY, peptide YY; TNF-α, tumor necrosis factor-α.

^aGLP-1 results were only available for 13 samples; none were from overweight gravidae and none had insufficient GWG.

Note: Bold indicates statistical significance.

When stratified by BMI, underweight gravidae demonstrated increased maternal serum leptin concentrations with increasing GWG (Kruskal–Wallis p = 0.003), normal weight gravidae had no change in any analyte concentrations by GWG, and gravidae with obesity demonstrated increased c-peptide (p = 0.017), insulin (p = 0.043), and leptin (p = 0.004); samples sizes are shown in Fig. 1. When stratified by GWG group, gravidae with insufficient GWG demonstrated no change in any analyte concentration by BMI group, gravidae with goal GWG persisted in demonstrating increased leptin concentrations with increasing BMI (p = 0.011), and gravidae with excess GWG demonstrated increased c-peptide (p = 0.048), GLP-1 (p = 0.026), and leptin (p = 0.013) with increasing BMI.

We also sought to evaluate inflammatory and metabolic markers by BMI and GWG as dichotomous variables. This analysis demonstrated that IL-6 was significantly lower among gravidae with obesity compared with those without obesity; no other inflammatory markers demonstrated statistically significant differences between the groups (Table 3). There were no statistically significant differences in inflammatory markers by GWG as a dichotomous variable. Regarding metabolic markers, leptin levels were statistically significantly higher among gravidae with obesity and were also higher among gravidae with excess GWG. Gravidae with excess GWG also had higher levels of c-peptide and insulin than gravidae without excess GWG (Fig. 2).

Table 3.

Cytokine and metabolic analytes by body mass index and gestational weight gain as dichotomous variables (Mann–Whitney’s U-test).

	BMI < 30	BMI ≥ 30	p-Value	No excess GWG	Excess GWG	p-Value
Analyte	n = 81	n = 39		n = 46	n = 74
Cytokines
IL-6	79.7 (212)	27.6 (58.4)	0.025	43.8 (62.8)	74.5 (221.9)	0.557
MCP-1	330 (380)	368 (536)	0.624	285 (280)	378 (506)	0.209
TNF-α	4.94 (3.77)	11.2 (35.5)	0.204	4.44 (1.83)	8.55 (26.0)	0.315
Metabolic analytes
Amylin	55.0 (40.5)	57.3 (51.5)	0.797	52.9 (50.3)	57.6 (40.2)	0.160
c-peptide	3,420 (2,278)	3,924 (2,806)	0.428	2,960 (2,259)	3,971 (2,517)	0.017
Ghrelin	25.6 (1.78)	26.0 (1.57)	0.118	25.5 (1.82)	25.8 (1.65)	0.660
GIP	57.3 (60.2)	56.4 (60.7)	0.695	47.7 (54.9)	62.8 (62.9)	0.104
GLP-1	5.14 (19.4)	8.28 (23.0)	0.557	1.61 (7.84)	8.99 (25.1)	0.064
Glucagon	28.3 (8.96)	28.0 (6.36)	0.838	26.9 (6.21)	29.0 (9.14)	0.239
Insulin	1,232 (1,362)	1,644 (2,333)	0.944	1,196 (1,872)	1,471 (1,654)	0.014
Leptin	25,240 (15,812)	35,789 (26,384)	0.028	21,733 (14,640)	32,980 (22,254)	<0.001
PP	73.8 (86.6)	63.4 (71.6)	0.692	66.7 (77.7)	72.8 (84.8)	0.741
PYY	41.1 (78.8)	48.7 (91.9)	0.930	50.6 (94.1)	39.1 (75.6)	0.741

Abbreviations: BMI, body mass index; GIP, gastric inhibitory polypeptide; GLP-1, glucagon-like peptide; GWG, gestational weight gain; IL-6, interleukin-6; MCP-1, monocyte chemoattractant protein-1; PP, pancreatic polypeptide; PYY, peptide YY; TNF-α, tumor necrosis factor-α. Bold indicates statistical significance.

Note: All analyte results are shown in pg/mL.

Fig. 2.

Leptin levels vary by both BMI and GWG. Stratification by BMI is shown with Kruskal-Wallis p-values shown. BMI, body mass index; GWG, gestational weight gain.

BMI as a continuous variable positively correlated with insulin (p = 0.048) and leptin (p = 0.002). GWG as a continuous variable positively correlated with TNF-α (p = 0.032) but no other analytes.

Investigating the analytes themselves, leptin levels were also found to be positively correlated with levels of c-peptide (p < 0.001), GIP (p = 0.015), glucagon (p = 0.002), and insulin (p = 0.003). TNF-α levels correlated with MCP-1 (p < 0.001). The c-peptide correlated with ghrelin (p = 0.013), GIP (p < 0.0001), glucagon (p = 0.015), insulin (p < 0.001), and PP (p = 0.006). Ghrelin was correlated with PP (p = 0.001) and PYY (p < 0.001). Insulin correlated with leptin and c-peptide (as above), but also amylin (p < 0.001), GIP (p < 0.001), PP (p = 0.001), and PYY (p = 0.009). Correlation p values are summarized in Table 4. In cases where the p values were significant, the Pearson r values remained low; the highest correlation observed was between insulin and amylin (r = 0.829).

Table 4.

Pearson’s product moment r and p values

	IL-6	MCP-1	TNF-α	Amylin	c-peptide	Ghrelin	GIP	GLP-1	Glucagon	Insulin	Leptin	PP
Cytokines
IL-6	–
MCP-1	0.476
p-Value	<0.001	–
TNF-α	0.071	0.663
p-Value	0.439	<0.001	–
Metabolic analytes
Amylin	−0.148	−0.140	−0.057
p-Value	0.108	0.128	0.536	–
c-peptide	−0.102	−0.047	0.000	0.661
p-Value	0.269	0.607	0.997	<0.001	–
Ghrelin	0.096	0.125	0.061	0.156	0.226
p-Value	0.295	0.173	0.511	0.088	0.013	–
GIP	−0.069	−0.080	−0.049	0.459	0.699	0.031
p-Value	0.454	0.388	0.592	<0.001	<0.001	0.736	–
GLP-1	−0.050	−0.050	0.000	0.053	0.167	−0.083	0.282
p-Value	0.591	0.585	0.999	0.568	0.068	0.365	0.002	–
Glucagon	0.009	−0.136	−0.037	0.244	0.221	0.090	0.170	0.098
p-Value	0.924	0.138	0.690	0.007	0.015	0.331	0.064	0.285	–
Insulin	−0.092	−0.091	−0.025	0.820	0.764	0.146	0.546	0.139	0.149
p-Value	0.316	0.324	0.789	<0.001	<0.001	0.112	<0.001	0.130	0.104	–
Leptin	−0.104	0.008	0.111	0.131	0.374	0.177	0.222	0.175	0.283	0.271
p-Value	0.260	0.928	0.227	0.155	<0.001	0.054	0.015	0.056	0.002	0.003	–
PP	−0.102	−0.087	−0.062	0.240	0.252	−0.293	0.422	0.296	0.668	0.306
p-Value	0.268	0.343	0.502	0.008	0.006	0.001	<0.001	0.001	0.668	0.001	0.857	–
PYY	−0.111	−0.116	−0.064	0.231	0.124	−0.460	0.393	0.278	−0.004	0.239
p-Value	0.225	0.206	0.484	0.011	0.178	<0.001	<0.001	0.002	0.966	0.009	0.440	<0.001

Abbreviations: GIP, gastric inhibitory polypeptide; GLP-1, glucagon-like peptide; GWG, gestational weight gain; IL-6, interleukin-6; MCP-1, monocyte chemoattractant protein-1; PP, pancreatic polypeptide; PYY, peptide YY; TNF-α, tumor necrosis factor-α.

Note: Bold indicates statistical significance.

Discussion

Here, we demonstrate that both BMI and GWG influence maternal circulating metabolic analytes, namely leptin, insulin, and c-peptide. These findings remain significant after stratifying by prepregnancy BMI (with the exception of gravidae with normal weight) with leptin demonstrating increased serum concentrations with GWG, even among underweight gravidae. Findings varied when stratifying by GWG, suggesting that both prepregnancy BMI and GWG play a role in maternal serum metabolic analyte levels at the time of delivery. Results also differed when results were analyzed as continuous variables, although this latter analysis is limited in properly accounting for GWG because appropriate GWG depends upon prepregnancy BMI.¹⁷

Our findings are consistent with prior studies which have demonstrated that obesity in general and during pregnancy is associated with aberrations in metabolic markers; however, prior studies also found that obesity is associated with alterations in markers of inflammation.^4-13 Adipose tissue produces circulating bioactive substances, such as leptin, and inflammatory markers, such as IL-6 and TNF-α in nonpregnant obesity.²⁰ Leptin itself modulates the innate immune response and stimulates the releases of proinflammatory cytokines and prostaglandins from maternal adipose tissue and also the placenta, specifically TNF-α, IL-6, IL-1 β, and PGE2.^{12, 13} Gravidae with obesity have demonstrated increased maternal circulating insulin, leptin, glucose levels, and lower adiponectin.^21-28 Regarding inflammatory markers, gravidae with obesity have increased CRP, ERP, and lower cortisol.^11,21,22 Why these findings differ from each other may depend upon the study population, the methods, the gestational age at sampling, or it may be that inflammatory or metabolic analytes may be modulated by sex hormone-binding globulin or other factors for which we are not accounting.²⁹ Gravidae with obesity have also been demonstrated to have alterations in microRNA, erythrocyte fatty acid composition, polyunsaturated fatty acids, adverse fatty acid profile and lipid profile, and placental inflammation, among other aberrations.^30-35

GWG, which exceeds the weight gain recommendations of the IOM,¹⁷ has been associated with multiple adverse outcomes.¹⁶ Here, our analysis was limited to term deliveries. Our findings are consistent with the findings of prior investigators who noted excess GWG to be associated with higher maternal insulin²² and leptin levels.²⁸ Prior investigators have also found excess GWG to be associated with adverse lipid profiles³⁵ and inflammatory markers.³⁶ One study uniquely found that polymorphisms in the leptin gene were associated with increased risk of excess GWG but not with leptin concentrations during pregnancy, but they did not directly evaluate for an association between GWG with leptin concentrations themselves.³⁷ However, that study does offer a potential mechanism to explain the relationship between leptin and GWG.

Prior studies have also demonstrated higher CRP in pregnancies with excess GWG.³⁶ We did not assess for this inflammatory biomarker, but the prior findings do suggest that there may be inflammatory changes with excess GWG. Other studies also found elevations in second trimester CRP in pregnancies with subsequent excess GWG,³⁸ but causality could not be determined. Interestingly, one prior study introduced a lifestyle intervention with the aim of reducing GWG; this intervention resulted in both lower GWG and CRP than controls, which suggests that nutrition and physical activity indeed play a role in GWG and CRP.

We did reliably observe changes in leptin by BMI and GWG. Leptin correlates with the percentage of body fat.³⁹ When gravidae with obesity gain excess weight during pregnancy, this weight gain is associated with accumulation of maternal adipose tissue rather than lean body mass.⁴⁰ This contrasts with weight that is gained within the guidelines, which is generally comprised of lean mass, although some prior studies are limited by alternate definitions of obesity.⁴¹ Leptin can also act as an inflammatory cytokine given its resemblance to IL-6.^12,42 Thus, while inflammatory analytes were generally not noted to be increased in gravidae with obesity or with excess GWG in this study, leptin’s role as an inflammatory mediator may still result in a pro-inflammatory milieu in gravidae with obesity and excess GWG.

Obesity and excess GWG have also been associated with preterm birth.⁴³ While this study exclusively examined maternal serum from pregnancies that proceeded to term, the findings support that both obesity and excess GWG (at term) result in increased levels of leptin, which as noted above can result in a pro-inflammatory milieu, which can contribute to mechanisms that result in preterm birth.²

Here, we investigated 13 inflammatory and metabolic analytes among maternal serum samples collected at term. Future studies may consider first characterizing these analytes across pregnancy and then compare deliveries that ultimately deliver at term to those delivering preterm.

Strengths of our study include the use of a validated perinatal database and biospecimen repository with inclusion of over 800 clinical variables and rigorous collection and storage techniques.¹⁵ Our study also tested a broad selection of analytes related to obesity via technology that has only been used in one other study during pregnancy, and the prior study evaluated the role of overweight, not obesity.²⁷ However, our study also has limitations. First, the prepregnancy weight for the majority of participants was self-reported. Self-reported prepregnancy weight is frequently used in the literature and the use of self-reported prepregnancy weight has been reported to be within a few pounds of actual prepregnancy weight.⁴⁴ The use of first trimester weight as a surrogate for instances when prepregnancy weight is unknown is also a reasonable substitution when the prepregnancy weight is unknown, and thus this was utilized in this study as described in the methods section. Second, this study’s use of our institution’s biospecimen repository of samples obtained at delivery limited our ability to assess the impact of GWG across the course of gestation or assess how these analytes varied with gestational age or stages of parturition. However, these may be investigated in future studies. Finally, metabolic biomarkers have been demonstrated to be altered by parturition⁴⁵; however, here, only a single blood sample was collected at admission for delivery, and thus any effect of the parturition process would apply to all samples.

Here, we demonstrate that both obesity and excess GWG result in aberrations in metabolic biomarkers, most reliably, leptin. These findings were seen even among gravidae whose prepregnancy BMI was underweight. Because leptin can contribute to inflammation, efforts should be focused on limiting GWG. Recent meta-analyses indicate that this is feasible, via dietary intake, exercise, or both.⁴⁶