Serum fibroblast growth factor 21: Lack of association with gestational diabetes and pregnancy outcomes

Abstract

BACKGROUND

The prevalence of gestational diabetes mellitus (GDM) has been increasing worldwide and is associated with multiple adverse pregnancy outcomes. Despite standard screening, some cases remain undiagnosed. Fibroblast growth factor 21 (FGF21) plays a role in modulating glucose metabolism. There is an ongoing controversy regarding the relevance of FGF21 to GDM.

AIM

To evaluate the association between early second trimester serum FGF21 levels and gestational diabetes, and its predictive potential for outcomes.

METHODS

This cross-sectional observational study was conducted at a tertiary medical center, Chiang Mai University, Thailand. It included 28 pregnant women diagnosed with GDM and 81 pregnant women with normal glucose status. Blood samples were collected according to the study schedule, and pregnancy outcomes were recorded meticulously. Descriptive analysis was employed to evaluate the data.

RESULTS

Most participants in our study had no risk factors for GDM (body mass index < 24 kg/m², no first-degree relatives with diabetes, no history of GDM), normal baseline glucose status (fasting glucose < 110 mg/dL), and no insulin resistance (homeostatic model assessment of insulin resistance < 2). There was a trend of increased FGF21 levels in the insulin-treated GDM group compared with dietary-treated GDM and non-GDM (73.58 pg/mL vs 62.94 pg/mL vs 63.59 pg/mL, respectively, P = 0.73). However, no significant association was found between FGF21 concentrations and pregnancy outcomes based on quintiles of FGF21 levels.

CONCLUSION

FGF21 was not associated with GDM or pregnancy outcomes; however, due to the small sample size, larger clinical trials with a diverse population are suggested to confirm these results.

Key Words: Adipokine; Diabetes mellitus; Fibroblast growth factor 21; Gestational diabetes; Pregnancy

Core Tip: Fibroblast growth factor 21 (FGF21) plays a role in glucose and lipid metabolism, prompting studies on its potential as a predictive marker for gestational diabetes mellitus (GDM). This investigation examined serum FGF21 levels in the early second trimester and their association with GDM. Although a trend toward higher FGF21 levels was noted in insulin-treated GDM patients, no significant association was found between FGF21 and GDM status nor pregnancy outcomes. Possible reasons include a small sample size, low body mass index among participants, and unexamined factors like physical activity. Further research is needed to clarify the role of FGF21 in GDM.

INTRODUCTION

Gestational diabetes mellitus (GDM) is one of the main contributors to adverse pregnancy outcomes. These consequences encompass an increased risk of pre-eclampsia[1], heightened rates of cesarean section deliveries[2], fetal macrosomia[3], maternal birth canal injuries[3], and an elevated susceptibility to obesity and diabetes in offspring[4]. Presently, the global prevalence of GDM is continuously increasing. Within the United States, this prevalence has escalated markedly, with figures ascending from 4.6% in 2006 to 8.2% in 2016, marking a 78% increase over a decade[5]. A systematic review and meta-analysis found that the pooled prevalence in Asia is 11.5%[6]. In the context of our institution, Maharaj Nakorn Chiang Mai Hospital, the incidence of GDM has exhibited a substantial upward trajectory over the past decade, surging from 5.88% in 2000 to 21.6% in 2023.

Fibroblast growth factor 21 (FGF21) is a hormone categorized as an adipokine, produced by various organs like adipose tissue, skeletal muscle, liver, and pancreas. Its primary function involves regulating glucose and lipid metabolism, contributing to increased insulin sensitivity, decreased triglyceride levels, and assisting in the maintenance of energy balance[7]. The significant involvement of FGF21 in modulating glucose metabolism within the body has led to a growing interest in its potential implications in GDM over recent years. Nevertheless, existing studies remain constrained, yielding conflicting results. A recent review underscored the necessity for additional research to elucidate whether FGF21 can serve as a viable biomarker for GDM[8]. The objective of this study was to assess the association between serum FGF21 levels in the early second trimester and GDM, with a focus on investigating its potential utility as a predictive biomarker for GDM.

MATERIALS AND METHODS

Study design and setting

This cross-sectional observational study was conducted at Maharaj Nakorn Chiang Mai Hospital, a tertiary medical center located in northern Thailand, spanning from January 2021 to October 2022. Ethical approval was obtained from the Institutional Review Board (No. 328/2563). The inclusion criteria comprised singleton pregnancies and attendance at antenatal care during the first trimester. Exclusion criteria encompassed preexisting diabetes, polycystic ovary syndrome, hypertension, renal or liver disease, active inflammation, cancer, or systemic disease, as well as the use of medications affecting glucose metabolism and loss to follow-up. Upon obtaining individual informed consent, demographic data including age, parity, pre-pregnancy weight, height, previous obstetric history, and any risk factors for GDM were meticulously recorded.

Blood testing regimen

The blood testing regimen consisted of two phases: (1) First, assessment of fasting glucose and fasting insulin during the initial trimester (gestational age less than 14 weeks); and (2) Second, evaluation of fasting glucose, fasting insulin, and FGF21 during the early stages of the second trimester, which coincided with the timing of the quadruple test (gestational age between 14 weeks and 18 weeks). Blood samples were obtained subsequent to an 8-hour to 12-hour fasting period. Fasting plasma glucose levels were quantified utilizing the glucose oxidase method with commercially available kits. Detection of fasting insulin was performed utilizing the Cobas e411 platform (Roche Diagnostics, Indianapolis, IN, United States) and employing a sandwich principle electrochemiluminescence immunoassay. Serum specimens designated for FGF21 analysis were aliquoted and preserved at -80 °C until processing. Quantification of serum FGF21 concentrations was accomplished using a commercial Human Quantikine enzyme-linked immunosorbent kit (Catalog, No. DF2100; R and D Systems, Minneapolis, MN, United States).

Management and monitoring

All participants underwent a universal oral glucose tolerance test (OGTT) between 24 weeks and 28 weeks of gestation, serving as the gold standard for GDM diagnosis in our study. The OGTT was conducted following a minimum 8-hour fasting period. A blood sample drawn at fasting was used to assess insulin and sugar levels. After that, blood samples at 1-hour, 2-hour, and 3-hour post-glucose ingestion were collected to assess sugar levels. Diagnostic thresholds were based on the criteria established by the National Diabetes Data Group, as follows: (1) Fasting glucose levels ≥ 105 mg/dL; (2) 1-hour levels ≥ 190 mg/dL; (3) 2-hour levels ≥ 165 mg/dL; and (4) 3-hour levels ≥ 145 mg/dL. A diagnosis of GDM was assigned if two or more values exceeded these thresholds.

Management of antenatal care in both groups, as well as blood glucose monitoring in the GDM group, adhered to standard protocols overseen by clinicians rather than researchers. Insulin resistance was evaluated through the homeostatic model assessment of insulin resistance (HOMA-IR) calculation, employing the following formula: Fasting glucose (mmol/L) × fasting insulin (μU/L)/22.5.

Outcomes assessment

The obstetric outcomes, including gestational age at delivery, birth weight, placental weight, and pre-eclampsia status, were recorded meticulously. Clinical and laboratory parameters of the patients were categorized into two groups based on the results of the OGTT, and subsequent comparative analyses were conducted between these groups.

Statistical analyses

The statistical analyses were conducted utilizing Stata/MP version 16.0 (StataCorp., LLC, College Station, TX, United States). Continuous variables were expressed as median (interquartile range), while categorical variables were presented as frequency (percentage). Student's t-test or analysis of variance was employed for variables exhibiting a normal distribution, whereas Mann-Whitney U test was utilized for quantitative variables displaying a non-normal distribution. The χ² test was employed to compare qualitative variables between groups. The statistical review of the study was performed by a biomedical statistician of our department.

RESULTS

Participant characteristics

This study encompassed 28 pregnant women diagnosed with GDM and 81 pregnant women with normal glucose status. Within the group diagnosed with GDM, only 3 out of 28 participants required insulin therapy. The remaining participants effectively managed their condition through dietary control. Baseline characteristics did not exhibit significant differences across the groups. The majority of participants in the study did not exhibit any identifiable risk factors for GDM, as outlined in Table 1.

Table 1 Baseline characteristic of the study population, n (%).

Characteristics	Non-GDM (n = 81)	Diet-treated GDM (n = 25)	Insulin-treated GDM (n = 3)	P value
Maternal age in years	29.37 ± 4.29	31.52 ± 4.67	30.67 ± 4.04	0.10
Parity				0.07
Nulliparity	43 (53.75)	9 (36.00)	0 (0.00)
Multiparity	38 (46.91)	16 (64.00)	3 (100.00)
Risk factor
Pre-pregnancy body mass index in kg/m2	21.88 (19.61-25.51)	22.00 (19.83-27.05)	20.00 (17.67-34.22)	0.96
Previous > 4000 g	1 (1.23)	0	0	> 0.99
Previous GDM (n = 8)	1 (1.23)	1 (4.00)	0	0.45
1st degree relative diabetes mellitus	15 (18.51)	4 (16.00)	2 (66.67)	0.15

Data are mean ± SD, n (%), or median (q₁-q₃). GDM: Gestational diabetes mellitus.

Pregnancy outcomes

Analysis of pregnancy outcomes, including maternal weight gain during pregnancy, gestational age at birth, prevalence of pre-eclampsia, and newborn Apgar scores, revealed no significant disparities, as shown in Table 2. It is noteworthy that neonatal birth weight and placental weight were observed to be the lowest in the cohort receiving insulin treatment for GDM; however, this discrepancy did not reach statistical significance when compared to the other groups. The total weight gain during pregnancy was lowest in the group managing GDM through dietary control and was at the lower limit of normal recommendation. However, these differences were not statistically significant among the study groups.

Table 2 Pregnancy outcomes, n (%).

Outcomes	Non-GDM (n = 81)	Diet-treated GDM (n = 25)	Insulin-treated GDM (n = 3)	P value
Gestational age at delivery in weeks	39 (38-39)	38 (38-39)	38 (33-40)	0.780
Birth weighting	3120 (2880-3320)	3050 (2970-3210)	2630 (1580-3090)	0.150
Placental weighting	580.16 ± 136.85	566.67 ± 119.34	410 ± 101.49	0.098
Fetal sex				> 0.990
Male	40 (49.38)	13 (52.00)	1 (33.33)
Female	41 (50.62)	12 (48.00)	2 (66.67)
Apgar score
5-minute	9 (8-9)	9 (8-9)	9 (9-9)	0.700
10-minute	10 (9-10)	10 (9-10)	10 (10-10)	0.800
Pre-eclampsia	7 (8.64)	2 (8.00)	0	> 0.990
Weight at delivery in kg	68.00 (61.30-76.00)	67.70 (58.00–82.50)	70.00 (69.60-80.70)	0.640
Weight gain during pregnancy in kg	13.00 (10.00-17.00)	11.40 (8.70-14.00)	13.60 (3.70-19.00)	0.480

Data are median (q₁-q₃), mean ± SD, or n (%). GDM: Gestational diabetes mellitus.

Glycemic and insulin levels

None of the participants displayed hyperglycemia nor insulin resistance throughout the study period (Table 3).

Table 3 Chemical profiles of the study population, n (%).

Chemical profiles	Non-GDM (n = 81)	Diet-treated GDM (n = 25)	Insulin-treated GDM (n = 3)	P value
1st visit
GA in weeks	12.00 (12.00-13.00)	12.00 (12.00-13.00)	12.00 (12.50-13.00)	0.800
Fasting glucose in mg/dL	77.0 (71-81)	78 (76-81)	98.5 (NA-NA)	0.190
Fasting insulin in μU/mL	7.74 (5.78–10.61)	8.64 (4.83–12.78)	11.71 (NA-NA)	0.480
HOMA-IR	1.52 (1.14–2.06)	1.65 (0.84-2.56)	4.49 (NA-NA)	0.320
HOMA-IR ≥ 2	10 (12.34)	4 (16.00)	1 (33.33)	0.320
2nd visit
GA in weeks	16 (16-17)	16 (16-17)	17 (17-17)	0.320
Fasting glucose in mg/dL	74.00 (68.50-77.00)	72.50 (69.50-77.50)	79.00 (72.00-110.00)	0.007
Fasting insulin in μU/mL	6.96 (5.10-11.82)	7.82 (5.73-10.05)	8.15 (4.85-15.98)	0.930
HOMA-IR	1.23 (0.91-2.29)	1.44 (1.04-1.74)	1.59 (0.86-4.34)	0.840
HOMA-IR ≥ 2	20 (24.69)	5 (20.00)	1 (33.33)	0.600
Fibroblast growth factor 21 in pg/mL	63.59 (54.56-75.57)	62.94 (56.97-69.26)	73.58 (57.28–88.85)	0.730
3rd visit
GA in weeks	25 (24-26)	25 (25-27)	26 (NA-NA)	0.230
Fasting glucose in mg/dL	72 (680-76)	74 (66-79)	71 (NA-NA)	0.350
Fasting insulin in μU/mL	9.92 (6.39-14.85)	9.55 (5.31-20.29)	13.2 (NA-NA)	0.870
HOMA-IR	1.69 (1.10–2.53)	1.79 (0.82–3.97)	2.57 (NA-NA)	0.680
HOMA-IR ≥ 2	25 (30.86)	8 (32.00)	1 (33.33)	0.660

Data are median (q₁-q₃) or n (%). GA: Gestational age; GDM: Gestational diabetes mellitus; HOMA-IR: Homeostatic model assessment of insulin resistance; NA: Not available.

Late first trimester (gestational age 12-13 weeks): Fasting glucose and fasting insulin levels were highest in the insulin-treated GDM group followed by the diet-treated GDM group and the non-GDM group. However, these differences did not reach statistical significance among the three groups (Table 3).

Early second trimester (gestational age 16-17 weeks): Fasting glucose levels in the insulin-treated GDM group were notably elevated and significantly higher compared to both the diet-treated GDM group and the non-GDM group. Fasting insulin levels were also highest in the insulin-treated GDM group, followed by the diet-treated GDM group and the non-GDM group, although statistical significance was not observed among the three groups (Table 3).

Early third trimester (gestational age 24-27 weeks): Fasting glucose and fasting insulin levels were highest in the insulin-treated GDM group, although differences were not statistically significant compared to the other groups (Table 3).

FGF21 levels

Regarding FGF21, which was of particular interest in its association with GDM, measurements were obtained during the early second trimester at a median gestational age of 16 weeks. There was a trend indicating increased FGF21 levels in the insulin-treated GDM group, although statistical significance was not observed compared to the other groups (Table 3).

Association with pregnancy outcomes: We categorized the results based on quintiles of FGF21 levels and found no significant association between FGF21 concentrations and pregnancy outcomes (Table 4).

Table 4 Fibroblast growth factor 21 by quintile and selected pregnancy outcomes.

Outcomes	Quintile of fibroblast growth factor 21 in pg/mL					P value
	1: 28.52-51.91	2: 53.4-59.23	3: 60.04–67.56	4: 68.24–76.70	5: 79.21-140.55
Maternal age in years	29.14 ± 4.56	30.19 ± 4.90	31.03 ± 4.82	29.05 ± 3.35	30.48 ± 4.17	0.810
Pre-pregnancy body mass index in kg/m2	22.97 (19.42–26.37)	23.00 (20.20-26.56)	20.50 (19.30-23.04)	21.80 (19.57-24.00)	22.50 (19.20-28.00)	0.370
Fasting glucose in mg/dL	77.50 (73.00-80.50)	79.50 (73.00-81.00)	76.50 (74.00-82.0)	77.50 (70.00-82.00)	78.50 (77.00-81.00)	0.990
Fasting insulin in μU/mL	8.57 (5.09-14.50)	9.93 (9.14-10.61)	7.76 (5.26-10.27)	5.08 (4.77-7.65)	8.89 (7.08-14.66)	0.200
Gestational diabetes mellitus	4 (19.05)	7 (33.33)	8 (38.10)	4 (19.05)	5 (23.81)	0.520
Pre-eclampsia	0 (0.00)	3 (14.29)	2 (9.52)	1 (4.76)	2 (9.52)	0.630
Birth weighting	3210 (295-3295)	3120 (2970-3380)	3060 (2820-3140)	3130 (2870-3335)	3025 (2890-3260)	0.400
Placental weighting	577.14 ± 94.01	608.89 ± 113.70	552.94 ± 97.52	555 ± 140.76	567.22 ± 196.79	0.740

Data are median (q₁-q₃), mean ± SD, or n (%).

DISCUSSION

This study did not show the association between FGF21 levels measured in the early second trimester and GDM status or any pregnancy outcomes, which might be attributed to the small sample size. Although FGF21 levels had a higher trend in the insulin-treated GDM group, this was not statistically significant. This observed trend may be explained by the physiological role of FGF21 in glucose metabolism and insulin sensitivity, suggesting that elevated levels could reflect an adaptive response to insulin therapy or the metabolic alterations associated with GDM. FGF21 is known to be involved in regulating lipid metabolism and may be upregulated in states of insulin resistance, which could explain the higher levels seen in the insulin-treated group[7]. While FGF21 plays a role in metabolic regulation, it may not directly influence pregnancy outcomes in the context of this study. Nevertheless, the trends observed warrant further investigation, particularly in larger cohorts, to explore the potential implications of FGF21 in GDM.

A recent review conducted by Ruszała et al[9] identified conflicting data regarding the serum concentrations of FGF21 levels in women diagnosed with GDM when compared to those in healthy pregnant women. Our study investigated the levels of FGF21 during the early second trimester, coinciding with the national Down Syndrome screening policy. The objective was to reduce the need for invasive procedures in patients. We hypothesized that if FGF21 levels during this period predicted GDM that it would allow for closer monitoring and management before the GDM screening typically conducted later in pregnancy.

Detecting GDM earlier is crucial, as diagnosis in the third trimester may be too late for optimal intervention. Previous research has demonstrated that fetal growth trajectories diverge from 12 weeks onward between pregnancies complicated by GDM and those with normal glucose levels, suggesting that maternal metabolism differences begin at this early stage[10,11]. Unfortunately, our investigation did not reveal a significant association between FGF21 levels and GDM status, consistent with prior findings. For instance, Gawlik et al[12] observed increased serum FGF21 levels among GDM cases compared to healthy pregnant women at 24-28 weeks gestation; however, this disparity did not attain statistical significance. Similarly, Mosavat et al[13], following adjustment for confounding factors via logistic regression, concluded that FGF21 was not associated with the risk of developing GDM. They noted lower FGF21 levels in the GDM group compared to healthy pregnant women. Additionally, studies by Stein et al[14] and Xu et al[15] also reported no significant differences in FGF21 levels between the GDM group and normal control pregnant women.

In contrast, a study revealed that levels of FGF21 were higher in the GDM group compared to the normal glucose tolerance group at gestational age 24-28 weeks[16]. Previous investigations have focused on the early pregnancy levels of FGF21 and its association with GDM risk. For instance, FGF21 levels were elevated in the GDM group with an odds ratio of 1.85 (95%CI: 1.13-3.01) at gestational age 6-15 weeks[17]. Notably, this positive association between FGF21 levels and GDM was observed primarily in individuals with overweight or obesity [body mass index (BMI) ≥ 24 kg/m²][17]. This observation may explain the absence of an association in our study, given that all participants had a low BMI.

However, Wang et al[18] reported significant differences in FGF21 levels at 14-21 weeks gestation between women with and without GDM, irrespective of BMI status. However, these differences were more pronounced among women with a BMI ≥ 28 kg/m² compared to those with a BMI < 24 kg/m². This finding suggests the potential for modification or interaction between BMI and FGF21 levels. Furthermore, it is speculated that elevated FGF21 levels may not directly contribute to insulin resistance but could be secondary to the effects of obesity on insulin and FGF21 sensitivity[19]. This notion is supported by the association between FGF21 and pre-pregnancy BMI[16]. The controversies might stem from variances in pathogenesis between obese and non-obese women with GDM[20].

Additionally, the mechanism of FGF21 resistance may vary among women with GDM who have different BMIs[18]. This observation, including in our study, underscores the need for further research across various BMI categories to better understand the underlying mechanisms of FGF21 response. The observed discrepancies could be explained by the hypothesis that elevated FGF21 levels act as a compensatory mechanism to counteract the metabolic stress induced by obesity-related pathological conditions, or alternatively, they could indicate FGF21 resistance due to compromised downstream signaling[21]. The acceleration of GDM development is facilitated by excessive body mass, initiating an inflammatory process within adipose tissue[22,23].

The majority of studies have reported elevated serum levels of FGF21 in women with GDM beyond 24 weeks of gestation. Bonakdaran et al[24] and Li et al[25] observed higher FGF21 levels in GDM women compared to pregnant women with normal glucose tolerance at 24-28 weeks of gestation. Similarly, Šimják et al[26] noted that FGF21 levels in women with GDM after delivery were higher than those in non-pregnant women. These findings suggest that the elevation of FGF21 levels in the GDM population could potentially represent a compensatory mechanism employed by the body to ameliorate this pathological state. Moreover, it is plausible that FGF21 contributes to the pathophysiological processes associated with GDM. Variations in FGF21 findings across studies, when categorized by gestational age, highlight the complexity of FGF21 regulation and secretion, which remains incompletely understood. Conducting longitudinal studies to evaluate FGF21 levels throughout different stages of gestation will help clarify the intricate mechanism underlying FGF21 response in GDM patients.

Gawlik et al[12] identified a correlation between FGF21 levels and serum insulin in type 2 diabetes mellitus, indicating a potential association with impaired glucose tolerance due to insulin resistance, as evidenced by the HOMA-IR. In contrast, FGF21 did not correlate with serum insulin or HOMA-IR in GDM pregnancies, as reported by Gawlik et al[12] and Bonakdaran et al[24], which is consistent with our study findings. The absence of such associations between FGF21 and GDM could be attributed to the predominance of low HOMA-IR values among our study participants. This notion is further supported by findings from Wu et al[17] and Zhang et al[19] wherein the relationship between circulating FGF21 and HOMA-IR lost significance following adjustment for BMI. However, several previous studies have indicated a positive correlation between FGF21 and insulin resistance. For example, Wang et al[27] and Mosavat et al[13] found a positive association between serum FGF21 concentrations and insulin resistance in women with GDM.

The role of FGF21 in the pathophysiology of GDM remains unclear, highlighting the need for further investigation into the relationship between this adipokine and GDM. Additional research is required before FGF21 can be used as a predictor of GDM. Furthermore, future studies should include participants with a wide range of BMI values and be conducted throughout the gestation period to provide deeper insights into the association between FGF21 levels and GDM.

There were several limitations in our study. The cross-sectional design and small sample size hindered our ability to detect differences between study groups effectively. Moreover, the lack of consideration for other known biomarkers for GDM and the absence of information regarding lifestyle factors or physical activity, which are associated with FGF21, further constrain the scope of our findings. Therefore, our results necessitate validation through prospective large-scale studies to better understand this association and the pathophysiological significance of FGF21 in GDM.

CONCLUSION

GDM is a multifactorial disorder influenced by the interplay of environmental and genetic factors, wherein various adipokines potentially contribute to its onset and progression. FGF21, with its involvement in diverse physiological pathways includes the enhancement of insulin sensitivity and the regulation of glucose and lipid metabolism. However, this study did not demonstrate a significant association between FGF21 levels and GDM status nor pregnancy outcomes. Several hypotheses may explain this finding, including: (1) The relatively small sample size may limit the statistical power to detect subtle differences; (2) The majority of participants had low BMI, which may not accurately reflect FGF21 levels; and (3) Other factors associated with FGF21, such as physical activity, other adipokines and myokines, were not explored. Therefore, additional information on FGF21 and its role in pregnancy is needed to determine whether it could emerge as a relevant factor in the pathogenesis of GDM.

ACKNOWLEDGEMENTS

The authors would like to thank Tongsong T for his mentorship and Sekararith R for her assistance in data tracking and collection.