Serum uric acid and the risk of gestational diabetes mellitus: a systematic review and meta-analysis

Abstract

Aims

Serum uric acid (SUA) is considered as a risk factor for gestational diabetes mellitus (GDM). However, current studies showed inconsistent results. This study aimed to explore the relationship between SUA levels and GDM risk.

Methods

Eligible studies were retrieved from PubMed, Web of Science, Embase, China National Knowledge Infrastructure, and Wanfang databases up to November 1, 2022. The pooled standardized mean difference (SMD) and 95% confidence interval (CI) were used to represent the difference in SUA levels between GDM women and controls. The combined odds ratios (OR) and 95% CI were applied to assess association between SUA levels and GDM risk. Subgroup analyses were conducted on study continents, design, and quality, detection time of SUA, and GDM diagnostic criteria.

Results

Totally 11 studies including five case-control and six cohort studies, in which 80,387 pregnant women with 9815 GDM were included. The overall meta-analysis showed that the mean SUA level in GDM group was significantly higher than in controls (SMD = 0.423, 95%CI = 0.019–0.826, p = .040, I² = 93%). Notably, pregnant women with elevated levels of SUA had a significantly increased risk of GDM (OR = 1.670, 95%CI = 1.184–2.356, p = .0035, I² = 95%). Furthermore, subgroup analysis performed on the detection time of SUA showed a significant difference in the association between SUA and GDM risk within different trimesters (1st trimester: OR = 3.978, 95%CI = 2.177–7.268; 1st to 2nd trimester: OR = 1.340, 95%CI = 1.078–1.667; p between subgroups <.01).

Conclusions

Elevated SUA was positively associated with GDM risk, particularly in the 1st trimester of pregnancy. Further studies with high quality are required to validate the findings of this study.

Keywords:

• Uric acid
• gestational diabetes mellitus
• systematic review
• meta-analysis

Introduction

Gestational diabetes mellitus (GDM) is one of the most common complications during pregnancy. It affects ∼14% of all pregnancies globally and the prevalence of GDM has continued to rise in recent years [1]. Women with GDM are more likely to experience a variety of adverse pregnancy outcomes not only for mothers but also for newborns, such as macrosomia, shoulder dystocia, cesarean section, and neonatal hypoglycemia [2]. Additionally, GDM was significantly associated with a disorder of maternal glucose metabolism and childhood adiposity in later life [3]. Therefore, it is valuable to identify potential risk factors of GDM, which could help to take early prevention to avoid developing GDM.

Serum uric acid (SUA) is the final product of the oxidation step of purine metabolism, and elevated SUA levels are the markers of pathological mechanisms of many diseases [4]. SUA is known as an indicator of metabolic syndrome in the general population, such as hypertension, diabetes, obesity, and dyslipidemia [5,6]. Previously study found that SUA was a strong and independent risk factor for type 2 diabetes in 10 years of follow-up in non-pregnant women [7]. Hyperuricemia could trigger endothelial dysfunction and insulin resistance [8,9]. In pregnant women, several studies have assessed the risk of developing GDM with different levels of SUA, but the results were inconsistent. Therefore, a systematic review and meta-analysis were conducted to determine the association of SUA level with GDM.

Materials and methods

Search strategy

This systematic review was carried out according to the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines [10], and registered with PROSPERO (ID: CRD42022375968). The electronic databases of PubMed, Web of Science and Embase, China National Knowledge Infrastructure, and Wanfang were searched by two investigators up to 1st November 2022. The following terms were used for the literature search: (‘uric acid’ OR ‘urate’ OR ‘hyperuricemia’ OR ‘UA’) AND (‘Diabetes, Gestational’ OR ‘Diabetes, Pregnancy-Induced’ OR ‘Diabetes, Pregnancy Induced’ OR ‘Pregnancy-Induced Diabetes’ OR ‘Gestational Diabetes’ OR ‘Diabetes Mellitus, Gestational’ OR ‘Gestational Diabetes Mellitus’ OR ‘GDM’). In addition, the possible relevant studies found in the references were also included. There were no language restrictions. All included studies were imported into Endnote X9 (Clarivate Analytics, Philadelphia, PA) for management, screen, and review.

Study selection criteria

A two-step selection procedure was conducted. Firstly, titles and abstracts were independently reviewed by two reviewers. Relevant studies were included if they meet the following criteria: (1) observational studies including cohort or case-control studies; (2) exploring the association between SUA and GDM; (3) SUA level was available in GDM group and normal glucose tolerance (NGT) group; (4) studies with SUA detected before the GDM diagnosis. Studies were excluded if (1) irrelevant records; (2) not human studies; (3) reviews, comments, or letters. Secondly, full texts were examined, and the study was excluded if the required information was incomplete to obtain the standardized mean difference (SMD), adjusted odds ratio (OR), and 95% confidence intervals (CIs). Any inconsistencies between the two reviewers were settled by a discussion with a third reviewer.

Data extraction

The standardized form was used to extract relevant information from each eligible study by two reviewers independently. The extracted data included the name of the first author, year of publication, country, study design, sample size, detection time of SUA, unit of SUA value, values of SUA, diagnostic methods of GDM, adjusted OR with 95% CI, and adjusted confounding factors.

Quality evaluation

The Newcastle-Ottawa Scale (NOS) was used for evaluating the quality of studies, according to three components including selection of participants, comparability of participants, and outcomes of interest [11]. Each study was awarded from zero to nine stars. The study with ≥7 stars was considered as a high-quality study, and 3–6 were considered moderate quality.

Statistical analysis

The SMD and 95% CI were calculated to measure the difference in serum uric acid level between GDM and control group. The level of uric acid presented as median with interquartile range was converted to the mean and SD [12,13]. Additionally, the pooled OR and its 95%CI were calculated to express the association between uric acid and the risk of GDM. The heterogeneities were considered if p-value <.05 in Cochran Q test or I² statistics >50%. The DerSimonian and Laird random effects model was used to calculate pooled effect size that presented as forest plots with 95%CI [14,15]. Otherwise, the inverse variance fixed effect model was adopted [16]. Subgroup analysis was carried out based on study design, continent, gestation weeks when uric acid was detected, GDM diagnostic criteria, and study quality. Sensitivity and cumulative analysis were performed to evaluate the reliability of this meta-analysis. Publication bias was evaluated by the Begger’s the Egger’s tests. All statistical analyses were conducted using R software (version 4.2.1).

Results

Study selection

Figure 1 demonstrates a flowchart of the study selection process. A total of 142 studies were initially retrieved, of which 122 studies were excluded after screening the titles and abstracts. Then 20 studies were reviewed for full-text assessment, nine of them were excluded: the detection time of SUA was later than GDM diagnosis in four studies [17–20], and the SUA value or risk estimates were unavailable in five studies [21–25]. Finally, 11 studies were included in this meta-analysis [26–36].

Figure 1. Flow chart of study selection.

Study characteristics

The characteristics of the 11 included studies are summarized in Table 1. These studies were published from 2009 to 2022, and were conducted in seven countries over four continents: China (n = 5) [27,30,33,35,36], the USA (n = 1) [34], Israel (n = 1) [31], Turkey (n = 1) [26], Chile (n = 1) [28], Egypt (n = 1) [32], and Bangladeshi (n = 1) [29]. The sample size was from 96 to 69,151 in these studies. A total of 80,387 pregnant women were involved in this meta-analysis, of which 9,815 were diagnosed as GDM. Of the 11 studies in this meta-analysis, five were case-control studies [26–30], and six were cohort studies [31–36]. The level of serum uric acid was detected in these included studies from 6 to 27 gestational weeks. Among them, four studies were detected in the first trimester [26,28,30,34], four studies in the first to the second trimester [27,31,32,35], and three studies in the second trimester [29,33,36]. There were six studies adopting the International Association of the Diabetes and Pregnancy Study Groups criteria (IADPSG) for GDM diagnosis [27–30,35,36]. Moreover, the SMD of serum uric acid between GDM group and control group could be calculated from nine studies [26–29,31–33,35,36], and the adjusted OR of SUA were obtained from seven articles [27,30,31,33–36]. According to Newcastle–Ottawa quality assessment, eight studies were considered as high quality [26,28,31–36], and the other three studies were of moderate quality [27,29,30] (Table S1).

Table 1. The characteristics of the included studies.

First author	Year	Country	Study design	Detection time of SUA	Sample size		GDM diagnoses	SUA level (unit)		Odds ratio (95%CI)
					GDM	Control		GDM	Control
Laughon S. K. [34]	2009	USA	Cohort study	6–11 GW	73	1497	Non-IADPSG	NR	NR	Q1: reference Q2: 1.62 (0.63–4.22) Q3: 2.35 (0.96–5.78) Q4: 3.25 (1.35–7.83)
Zhou J. [33]	2012	China	Cohort study	20 GW	100	900	Non-IADPSG	231.3 ± 62.7 μmol/L	217.9 ± 50.3 μmol/L	Non-hyperuricemia (≤420 μmol/L): reference Hyperuricemia (>420 μmol/L): 2.34 (1.44–3.83)
Wolak T. [31]	2012	Israel	Cohort study	<20 GW	417	5090	Non-IADPSG	3.4 ± 1.0mEq/L	3.2 ± 0.9mEq/L	1.3 (1.1–1.5)
Maged A. M. [32]	2014	Egypt	Cohort study	<15 GW	27	242	Non-IADPSG	3.43 ± 0.61 mg/dl	3.29 ± 0.69 mg/dl	NR
Şahin Aker S. [26]	2016	Turkey	Case-control study	<14 GW	66	202	Non-IADPSG	5.95 ± 0.97 mg/dl	3.76 ± 1.07 mg/dl	NR
Correa P. J. [28]	2019	Chile	Case-control study	<14 GW	16	80	IADPSG	2.98 ± 0.75 mg/dl	2.73 ± 0.74 mg/dl	NR
Mishu F. A. [29]	2019	Bangladeshi	Case-control study	15–27 GW	86	86	IADPSG	4.47 ± 0.42 mg/dl	4.43 ± 0.63 mg/dl	NR
Li Y. [36]	2020	China	Cohort study	16–18 GW	396	1373	IADPSG	216.45 ± 46.07 μmol/L	209.36 ± 42.87 μmol/L	Q1: reference Q2: 1.195 (0.840–1.699) Q3: 1.111 (0.777–1.588) Q4: 1.463 (1.034–2.070)
Li J. [27]	2021	China	Case-control study	11–17 GW	80	317	IADPSG	220.50 ± 62.00 μmol/L	204.00 ± 60.00 μmol/L	1.003 (0.997–1.009)
Zhao Y. [35]	2022	China	Cohort study	<24 GW	8494^a	60657^a	IADPSG	208.05 ± 40.78^b	199.05 ± 37.81^b	Q1: reference Q2: 1.11 (1.05–1.22) Q3: 1.27 (1.17–1.37) Q4: 1.37 (1.27–1.48)
Hu T. [30]	2022	China	Case-control study	6–14 GW	60	90	IADPSG	NR	NR	4.76 (2.08–10.9)

CI: confidence interval; GDM: gestational diabetes mellitus; GW: gestational week; IADPSG: the International Association of the Diabetes and Pregnancy Study Groups criteria; NR: not reported; Q1–Q4: quartile 1–4; SUA: serum uric acid.

^aThis number was not directly reported in the original article and was calculated according to the number of each subgroup and the proportion.

^bThese data were obtained according to the calculation using the median and range which were given in the original article.

Different levels of SUA between GDM and NGT

The pooled results of SMD and 95%CI in nine studies were shown in Figure 2(A). The SMD between GDM and NGT was calculated in 78,629 pregnant women and 9682 cases with GDM, which was ranging from 0.074 to 2.087. The combined results showed that the SUA level was significantly higher in GDM group than in NGT group (SMD = 0.423, 95%CI = 0.019–0.826, p = .040). There was statistically significant heterogeneity qualified by heterogeneity test (I² = 94%, p < .01). Sensitivity analysis was conducted by calculating the SMD and 95% CI after each study omitted at a time. The pooled SMD was from 0.232 to 0.466 and there were no significantly diverged after removing each study (Figure S1(A)). The potential publication bias was evaluated by Egger’s test (p = .054) and Begger’s test (p = .095), showing there was no evidence of publication bias (Figure S2(A)).

Figure 2. Forest plots of the SMD in the levels of SUA between GDM and NGT Groups; (A) results of overall meta-analysis; subgroup analysis according to (B) study continents, (C) study design, (D) detection time of SUA, (E) GDM diagnostic approach, and (F) methodological quality. CI: confidence interval; GDM: gestational diabetes mellitus; IADPSG: the International Association of the Diabetes and Pregnancy Study Groups criteria; SD: standard deviation; SMD: standardized mean difference; SUA: serum uric acid.

Subgroup analyses were conducted on continents, study design, detection time of SUA, GDM diagnostic approach, and methodological quality, shown in Figure 2. The levels of SUA were not significantly different between GDM and NGT groups in different continents (Africa: SMD = 0.205, 95%CI = −0.194–0.603; Asia: SMD = 0.465, 95%CI = −0.050–0.979; South America: SMD = 0.334, 95%CI = −0.205–0.873; p among subgroups = .732) (Figure 2(B)), and different study design (case-control study: SMD = 0.694, 95%CI = −0.230–1.619; cohort study: SMD = 0.232, 95%CI = 0.211–0.254; p among subgroups = .328) (Figure 2(C)). Subgroup analysis on the detection time of SUA was performed by classifying the gestational weeks into the 1st trimester (≤14 weeks of gestation), and 1st–2nd trimester (≤27 weeks of gestation). SUA levels were not significantly different among the different detection times (1st trimester: SMD = 1.224, 95%CI = −0.493–2.941; 1st to 2nd trimester: SMD = 0.232, 95%CI = 0.210–0.253; p among subgroups = .257) (Figure 2(D)). In addition, there was no significant difference in SUA level between GDM and NGT groups in different GDM diagnostic criteria (IADPSG: SMD = 0.224, 95%CI = 0.180–0.268; non-IADPSG: SMD = 0.689, 95%CI = −0.217–1.594; p among subgroups = .315) (Figure 2(E)), and methodological quality (high-quality studies: SMD = 0.495, 95%CI = −0.020–1.010; moderate quality studies: SMD = 0.192, 95%CI = 0.002–0.383; p among subgroups = .280) (Figure 2(F)).

Association between SUA levels and GDM risk

The pooled results of OR and 95%CI in seven studies were shown in Figure 3(A). The association between SUA and GDM was evaluated in 79,544 pregnant women and 9620 cases with GDM, of which OR was ranging from 1.003 to 4.760. The summary OR of increased SUA for GDM risk was 1.670 (95%CI = 1.184–2.356, p = .0035), with substantial heterogeneity in these included studies (I² = 95%, p < .01). Sensitivity analysis showed that the pooled OR remained stable between 1.432 and 1.840, after removing each study (Figure S1(B)). The publication bias was observed by Egger’s test (p = .007) but not Begg’s test (p = .652) (Figure S2(B)).

Figure 3. Forest plots of the association between SUA levels and GDM Risks. (A) results of overall meta-analysis; subgroup analysis according to (B) study continents, (C) study design, (D) detection time of SUA, (E) GDM diagnostic approach, and (F) methodological quality. CI: confidence interval; GDM: gestational diabetes mellitus; IADPSG: the International Association of the Diabetes and Pregnancy Study Groups criteria; OR: odds ratio; SUA: serum uric acid.

Subgroup analysis was performed on continents, study design, detection time of SUA, GDM diagnostic approach, and methodological quality (Figure 3). The association between SUA and GDM was similar in different continents (Asia: OR = 1.550, 95%CI = 1.115–2.154; North America: OR = 3.250, 95%CI = 1.350–7.827; p among subgroups = .122) (Figure 3(B)), study design (cohort study: OR = 1.380, 95%CI = 1.292–1.475; case-control study: OR = 2.063, 95%CI = 0.450–9.452; p among subgroups = .605) (Figure 3(C)), GDM diagnostic criteria (IADPSG: OR = 1.570, 95%CI = 0.928–2.659; non-IADPSG: OR = 1.920, 95%CI = 1.122–3.283; p among subgroups = .601) (Figure 3(D)), and methodological quality (high-quality studies: OR = 1.380, 95%CI = 1.292–1.475; moderate quality studies: OR = 2.063, 95%CI = 0.450–9.452; p among subgroups = .605) (Figure 3(E)). A significant difference was found in the association between SUA and GDM within different detection times of SUA (1st trimester: OR = 3.978, 95%CI = 2.177–7.268; 1st to 2nd trimester: OR = 1.340, 95%CI = 1.078–1.667; p among subgroups <.01) (Figure 3(F)).

Discussion

To the best of our knowledge, there is no comprehensive evaluation of the association between the levels of SUA and the risk of GDM, and this is the first meta-analysis to identify a prospective association between them. This systematic review included 11 observational studies and involved 80,387 pregnant women, of which 9,815 were diagnosed with GDM. The overall results revealed that the higher levels of SUA were obviously found in GDM group compared to NGT group. More importantly, higher SUA levels were positively associated with an increased risk of GDM. In particular, this association was significantly stronger in the 1st trimester of pregnancy. These results were stable after sensitivity analysis, and also persist across subgroup analysis of study area, study design, GDM diagnostic criteria, and study quality, indicating that SUA may be involved in the development of GDM and used as a potential earlier biomarker for screening or predicting GDM.

Although several studies have reported that higher levels of SUA were associated with insulin resistance and increased risk of Type 2 diabetes among the general population, the associations of SUA with GDM risk were inconsistent and relatively limited. This current meta-analysis found that the mean level of SUA was significantly higher in GDM pregnant women than that in NGT pregnant women, which is similar to the study showing that individuals with diabetes often had a higher level of SUA, compared to the general population without diabetes [37]. In addition, our pooled OR of SUA levels for GDM was 1.670 (95%CI = 1.184–2.356, p = .0035), suggesting that the elevated SUA levels in the first to second trimester could significantly increase the risk of GDM. This is consistent with the previous meta-analysis conducted for the general population of 42834 participants and 3305 developed type 2 diabetes during follow-up, which included 11 cohort studies and showed that each 1 mg/dl increase in SUA, the risk of type 2 diabetes was increased 1.17-fold [38]. Moreover, this association was found not only between SUA levels and incidence of type 2 diabetes but also between SUA levels and risk of impaired fasting glucose (IFG) in another meta-analysis of 12 cohort studies involving 62834 participants and 6340 cases of IFG and type 2 diabetes [39]. This study showed that the pooled adjusted relative risk (RR) (95%CI) of IFG and T2DM for the highest level of SUA was 1.54 (1.41–1.68) compared to the lowest level of SUA. Furthermore, the higher levels of SUA may contribute to the diabetes-related complications. There has been some meta-analysis showed that elevated SUA was associated with the increased risk of diabetic kidney disease and all-cause mortality and stroke among type 2 diabetes population [40,41]. Thus, current research suggested that the effect of SUA on glucose metabolism is relatively consistent, regardless of pregnancy stage.

In the subgroup analysis by the detection of SUA, the association between SUA levels and GDM risk significantly differ in different trimesters: the stronger association was found in the 1st trimester than 1st–2nd trimester, indicating that the increase of SUA levels in the 1st trimester of pregnancy is more likely to experience GDM. The levels of SUA change in different stages of normal pregnancy. A previous study reported that the concentration of uric acid was reduced significantly by the 8th gestational week, compared with pre-pregnancy values, and these decreased values were maintained until 24th gestational week [42]. A similar result was obtained in Boyle’s study, showing that the mean SUA level was significantly lower during early and middle pregnancy in healthy pregnant women than in non-pregnant age-matched participants, and was not significantly different in late pregnancy compared to the controls [43]. Also, the recent study further observed that SUA levels decreased between 25 to 35% at the early stage of pregnancy in uncomplicated pregnant women compared to non-pregnant women, then the levels of SUA were increased slightly and arrived at a similar level to non-pregnant women at the end of gestation [44]. The reduction of SUA level in early and middle pregnancy may be caused by the pregnancy induced blood volume expansion, renal blood flow increases, and the estrogen’s uricosuric action [45], then with the fetal production increased, the binding to albumin decreased, and the renal clearance declined, SUA levels increased during the late pregnancy [42]. Therefore, elevated SUA level might be considered as a biological indicator to predict GDM risk, particularly in the 1st trimester of pregnancy.

The mechanisms of elevated SUA levels increasing the risk of GDM incidence are still needed to be interpreted. High level of SUA could induce inflammation and oxidative stress [46], and fructose-induced hyperuricemia was believed to mediate fructose-induced insulin resistance [47], leading to insulin resistance and glucose uptake decreasing [48]. Moreover, animal studies demonstrated that elevated SUA could affect β cell glycolysis and impair insulin secretion [49], which might result in GDM. Furthermore, previous study showed that glucose uptake could be enhanced, which depends on nitric oxide by increasing blood flow to adipocytes and muscle cells [50]. The accumulation of SUA could mediate endothelial cell dysfunction and decrease nitric oxide production, which might reduce the production of endothelium dependent vascular relaxation [51].

This is the first meta-analysis to summarize SUA levels between GDM and NGT pregnant women and to assess the effect of elevated SUA levels on GDM risk. Our study provides a detailed and comprehensive analysis and suggests that SUA plays an important role in the development of GDM and could be considered as a potential earlier biomarker for GDM. However, there are limitations to our study. First, heterogeneity across the original studies was observed. To reduce the effect of heterogeneity in this meta-analysis, we conducted random effect modelings to estimate the combined effect sizes. In addition, we performed the subgroup analysis based on study continents, study design, detection time of SUA, GDM diagnostic approach, and methodological quality, and calculated SMD between the GDM and NGT pregnant women. Furthermore, sensitivity analysis was also adopted to examine the stability of the results. Second, due to the individual data of SUA level was unavailable in each original study, there might be certain deviations for the subgroup analysis using the mean level of SUA. Third, five case-control studies were included in this meta-analysis and only one of them was nested case-control study, so there might have unavoidable biases. Fourth, there is publication bias existed in the analysis of the association between SUA levels and GDM risk. This may be caused by the exclusion of original studies which include insufficient information to extract data and calculate the effect sizes. Thus, more well-designed and prospective studies are required to validate the effect of SUA levels on GDM risk.

Conclusion

In conclusion, the findings of this systematic review and meta-analysis showed that SUA levels in GDM women are significantly higher than those in pregnant women with normal glucose tolerance. Notably, elevated levels of SUA are positively associated with an increased risk of GDM, particularly in the 1st trimester of pregnancy. Further studies with high quality are required to validate the findings of this study.

Authors’ contributions

SS and EZ searched and screened the articles, extracted and analyzed data, and prepared the manuscript. SG and YZ provided the related methodology. JL, SX, and WY contributed to revising the manuscript. RL and CY designed this study and interpreted the results. All authors reviewed and approved the final version of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

This meta-analysis was based on published articles. All the data used in this study was available in this current article. The raw data can be found in the original reviewed articles.

Additional information

Funding

This work was supported by the National Key Research and Development Program of China (2016YFC1000101); Beijing Obstetrics and Gynecology Hospital, Capital Medical University, Beijing Maternal and Child Health Care Hospital ‘Excellent Youth’ Plan Special Funds (No. YQRC201907); and Beijing Hospitals Authority Innovation Studio of Young Staff Funding Support (202130).