Gestational diabetes mellitus (GDM) is an inflammatory condition that involves unbalanced cytokine production. We carried out a systematic review on the relationship between GDM and maternal circulating levels of cytokines in the 2nd/3rd trimesters.
Method of Study
Three electronic databases (MEDLINE, EMBASE and LILACS), were searched. Duplicate study selection, extraction and quality assessment was performed.
Twenty-two studies with 1982 participants reporting levels of 9 cytokines (IL-1B, IL-2, IL-6, IL-10, IL-13, IL-18, IFN-G, TGF-B and TNF-A) were included. Most studies differed considerably in selection criteria, sampling and assay methods and in reporting their results. Consequently, only two studies could be pooled: TNF-A concentration was slightly higher in GDM than in control patients, although not significant (WMD = 0.45, 95% CI −0.34–1.23).
New studies with well-defined, more homogeneous methodological parameters are needed to detect whether there are significant differences in circulating levels of cytokines in patients with GDM.
Gestational diabetes mellitus (GDM), defined as carbohydrate intolerance with onset or first recognition during pregnancy, is estimated to affect 1–22% of all pregnancies, depending on population characteristics and diagnostic criteria used. Worldwide, the prevalence of GDM has been steadily increasing over the last 20 years, in part due to the obesity epidemic resulting from lifestyle changes and also due to the increasing number of women delaying pregnancy until later in life. With the adoption of the new diagnostic criteria proposed by the IADPSG, the prevalence of GDM is expected to increase fourfold in the coming years.
Besides increased obstetric and perinatal morbidity, GDM is associated with long-term consequences for the mother and her infant including the development of metabolic syndrome, type 2 diabetes (T2DM) and cardiovascular disease.[1, 6] Despite some progress in the field, the pathogenesis and physiopathology of GDM are not yet completely clear.
In recent years, the role of the inflammatory system in the pathogenesis of T2DM and GDM has been increasingly investigated.[8, 9] Cytokines, a group of proteins that are expressed by several cell types, act as immune mediators and regulators. Depending on the period of pregnancy, a predominant inflammatory profile defined by increased production of Th1 cytokines (such as interferon-gamma (IFN-G) and tumor necrosis factor-alpha (TNF-A)), may compromise the normal development of the concept, while an anti-inflammatory pattern characterized by increased production of Th2 cytokines (such as IL-4, IL-6 and IL-10) seems to favor a normal pregnancy outcomes.[11-14]
Insulin resistance has been associated with abnormal secretion of pro-inflammatory cytokines such as TNF-A and interleukin (IL)-6 and decreased production of anti-inflammatory mediators such as IL-4 and IL-10.[7, 15] Despite some controversies regarding specific cytokine levels, T2DM is currently regarded as a chronic inflammatory disease.
Due to the similarity between T2DM and GDM and the clear relationship between T2DM and inflammation, it has been hypothesized that inflammation could be also implicated in the pathophysiology of GDM. Several studies have investigated the inflammatory response and cytokine production of women with GDM compared with healthy pregnant controls. While some authors report increased TNF-A levels in GDM,[16-22] others do not confirm this association.[23-28] Similar controversies also exist regarding IL-10, with study showing lower levels of this anti-inflammatory cytokine in patients with GDM whereas others do not confirm these findings.[23, 25, 27]
The observed controversies could be related to differences in the characteristics of the participants, including gestational age and severity of the disease, to different types of sample (serum, plasma or culture supernatant), to the use of different assay methods to measure the concentration of cytokines (i.e., ELISA, chemiluminescent immunoassay or immunoradiometric assay), and also to the lack of adjustment for maternal body mass index (BMI) as well as other important factors such as smoking and ethnicity that affect the production of cytokines.[30-32]
To clarify the role of cytokines in the physiopathology of GDM, it is essential to first map out the existing studies on this topic and analyze their findings. To the best of our knowledge, up to the present there have been no previous systematic reviews of the literature that retrieved, analyzed and synthesized the findings of studies on cytokine levels in patients with GDM compared with healthy pregnant women. This motivated us to perform such a review, on cytokine levels in patients with GDM.
Material and methods
This review followed the recommendations for systematic reviews of observational studies.
Selection Criteria for Including Studies
Types of studies
Studies that assessed the following cytokines in the peripheral blood (serum/plasma/culture supernatant) of women with GDM and of healthy controls were eligible for inclusion in the review: IL-1B, IL-2, IL-6, IL-10, IL-13, IL-18, IFN-G, transforming growth factor beta (TGF-B) and TNF-A. Studies using any type of assay method were considered eligible for inclusion. We accepted all types of study designs (cross-sectional, case–controls, nested case–controls and cohort studies). Studies without a healthy control group, that is, reporting cytokine levels only in patients with GDM (case series), were excluded.
Patients with GDM and healthy control women in the 2nd or 3rd trimesters of pregnancy were included in the review. Prediction studies that analyzed cytokines in women before the onset of GDM (e.g., in the 1st or early 2nd trimester of pregnancy) as potential biomarkers for the future development of the disease were excluded from this review. Any criteria used by the authors for the diagnosis of GDM were accepted.
Studies were ineligible if any of the following applied: (i) studies that included pregnant women with all types of DM (type 1, type 2 or GDM) but did not provide separate data on GDM, (ii) exclusively fetal and or placental tissue studies (e.g., fetal biopsy or cord blood, placental biopsies), (iii) animal model studies, (iv) studies that did not report the number of cases and or controls, (v) studies that did not provide concentrations of cytokines in cases and/or controls, (vi) editorials, comments or review articles without original data, (vii) proteomic studies, (viii) tissue-based studies and mRNA expression studies.
Search strategy and process of study selection
We searched three electronic databases (MEDLINE, EMBASE and LILACS), for articles published from inception up to October 2012. Studies published in English, Spanish, Portuguese, French or Italian were included. The search terms were: ‘gestational diabetes’ combined with ‘cytokines’ OR the names of the specific molecules included in the review, that is, ‘IL-1B’, OR ‘IL-2’ OR ‘IL-6’ OR ‘IL-10’ OR ‘IL-13’ OR ‘IL-18’ OR ‘IFN-G’ OR ‘TGF-B’ OR ‘TNF-A’ and their synonyms, adapted to each specific database. Detailed search strategy can be obtained from authors upon request. The reference lists of all articles selected for full-text reading were reviewed for additional potentially eligible studies. All retrieved references were downloaded into an electronic reference manager database, and duplicates were removed. Based on the aforementioned selection criteria, the titles and abstracts of retrieved references were screened for potential inclusion in the review. The full-text articles of selected references were obtained, and those fulfilling the selection criteria were extracted and included in the final review.
Two independent reviewers conducted in duplicate the whole process of screening, full-text reading and study selection. Discrepancies were discussed until consensus was reached.
A data extraction form was created to collect the following information from each article included in this review: study design and setting, inclusion and exclusion criteria, participant characteristics, total number of participants and of cases and of controls, diagnostic criteria for GDM, severity of GDM (need for insulin), gestational age at sampling, description of sample collection and storage, description of laboratory method used to assay cytokine concentrations, results in cases and controls, adjustment for potential confounders. Data were extracted by two independent investigators and compared. Discrepancies were discussed until consensus was reached, and a final data extraction form was obtained for each study.
Quality assessment of studies
The reviewers used a defined set of parameters created specifically for this review based on the QUADAS tool to assess and grade the quality of included studies. The following parameters were assessed and graded for each study: (i) loss of follow-up of participants, (ii) description of population characteristics and inclusion/exclusion criteria, (iii) description of sample collection, handling and laboratory procedures, and (iv) quality of the results presented (sample size calculation and/or statistical power of the study, adjustments for potential confounders or effect modifiers). Each of the four aforementioned parameters was graded as good, regular or poor according to the content, clarity and details of the information presented by the authors (Fig. 1).
Data synthesis and analysis
All data extracted from each study were grouped in an electronic spreadsheet. Population characteristics were presented descriptively as range, mean and standard deviation, and absolute and relative frequency (percentage). The absolute difference between the mean cytokine concentrations in the GDM and control participants in each study was calculated. Weighted mean difference (WMD) analyses were performed for studies that used the same scale. Data presented as multiples of median were not converted. Continuous data were pooled, and meta-analysis combining the weighted mean differences across studies was performed using the RevMan 4.2 software (the Nordic Cochrane Centre, Copenhagen, Denmark). The I2 statistic was used to assess heterogeneity between studies. In the absence of significant heterogeneity, results were pooled using a fixed-effect model. If substantial heterogeneity was detected (I2 > 50%), a random-effects model was used.[36-38]
The electronic search yielded 847 citations, which were reduced to 725 after eliminating duplicates. After screening titles and abstracts, 31 references were selected for full-text reading[16-29, 39-55] and 23 citations reporting on 22 studies were included in this review (Fig. 2).
The 22 studies reported 9 different cytokines measured in a total of 1,982 women, 1,027 of whom had GDM. Table 1 presents the main characteristics of these studies. All were published in the last 10 years and were case–controls. The number of participants ranged from 15 to 250 (mean = 96), although most of the studies (15/22) included <100 women. The number of patients with GDM per study ranged from 5 to 150 (mean = 47) with only two studies including >100 cases. The GDM diagnostic criteria proposed by the World Health Organization were used by 45% of the studies. Disease severity varied widely among the patients with GDM: 31.8% of the studies included patients treated with diet and/or insulin, 27.3% included only patients treated with diet, 18.2% included only patients treated with insulin, and 22.7% of the studies did not provide details on the treatment of their patients with GDM. Variables that can affect cytokine levels were reported in most of the studies, to some extent. Twenty of the studies (90.9%) provided information on the participants′ BMI, seven informed their ethnicity,[19, 23, 25, 27, 42, 43, 46] and seven informed their smoking status.[16, 18, 23, 24, 29, 42, 46] Most of the studies were judged to be of moderate or high quality in all domains assessed. The worse domain was quality of results: almost one-third of the studies provided no information on sample size calculation and/or power and did not adjust their results for potential confounders/effect modifiers (Fig. 1).
Table 1. Main characteristics of 22 studies on cytokine levels in patients with GDM
BMI Mean (S.D.) Range
ADA, American Diabetes Association; C, control group; GDM, gestational diabetes mellitus; NDDG, National Diabetes Data Group; NI, no information; NS, not significant; WHO, World Health Organization; CDA: Canadian Diabetes Association.
Most of the studies assessed cytokine levels in serum (11/22)[16, 18, 21, 24, 26, 28, 39-42, 46] or plasma (9/22)[17, 19, 20, 22, 25, 27, 29, 44, 45] samples. Commercial immunoenzymatic assay (ELISA) kits were used for cytokine analyses in almost all studies (18/22). Only seven studies provided description of sample collection (i.e., tubes, anticoagulant) and storage (temperature, time) details.[23-28, 44]
Nine cytokines were analyzed in the 22 included studies, TNF-A was the most frequently investigated cytokine[16-28, 40, 41] followed by IL-6[17, 23, 27, 29, 39, 42, 45] and IL-10.[23, 25, 27, 29] Almost 70% of the studies assessed only one cytokine.[16, 18-22, 24, 26, 28, 40-42, 44-46] Table 2 presents cytokine concentrations in women with and without GDM. A summary of findings of the 22 studies on the cytokine levels in patients with GDM compared with healthy controls is presented on Table 3.
Table 2. Cytokine concentrations in women with and without GDM
Although 15 studies evaluated TNF-A, only two[16, 28] were similar enough to allow pooling of their results into a meta-analysis. Eleven studies[17-20, 22-27, 41] could not be pooled because of differences in participant selection criteria or gestational age at sampling or type of sample or assay methods. Two studies[21, 40] did not provide essential details to allow pooling of their results. According to the data from the two studies included in the meta-analysis, the mean differences in TNF-A levels were slightly higher in patients with GDM than in controls, but this did not reach statistical significance (Fig. 3).
Gestational diabetes mellitus is an inflammatory condition and as such, cytokines have been implicated in its physiopathology. In the last decade, there have been numerous publications on cytokines and GDM; however, the findings are controversial, and this relationship is not yet completely understood. We performed this systematic review to synthesize the findings of these studies.
Our search identified 22 publications that met our selection criteria. However, despite this relatively large number of studies, due to methodological differences only two[16, 28] of them could be pooled. According to this meta-analysis, patients with GDM in the late 2nd/early 3rd trimester of pregnancy have slightly higher, albeit non-significant, TNF-A serum levels than healthy controls at the same gestational age. In accordance with this finding, over 70% of the 15 studies on TNF-A reported an association between this cytokine and GDM.[16-22, 26, 28, 40, 41]
This finding is not unexpected, as TNF-A is an inflammatory cytokine that plays a central role in the development of T2DM.[56, 57] Probably due to this fact, TNF-A was the cytokine with the largest number of studies in this review about studies on GDM. However, due to differences in participant selection criteria, gestational age at sampling type of sample, assay method and data reporting, only two studies[16, 28] could actually be included in the meta-analysis.
The second most frequently studied cytokine was IL-6, with 7 studies: 4 showing similar levels and 3 showing higher levels of this cytokine in patients with GDM compared with healthy controls. However, it should be noted that these 7 studies analyzed different types of samples (serum, plasma and culture supernatant), collected at different gestational ages ranging from 25.6 to 38.2 weeks.[17, 23, 27, 29, 39, 42, 45] As IL-6 has inflammatory properties and may induce insulin resistance, it was expected that patients with GDM would have higher levels of this cytokine than healthy pregnant women. Moreover, experimental studies have shown that high glucose concentrations stimulate IL-6 production. In addition, high IL-6 concentrations have been associated with obesity, metabolic syndrome and type 2DM.[56, 59] The unexpected results in four of the studies could in part be attributed to methodological differences, including differences in the selection criteria of the participants, gestational age at sampling and lack of adjustments for confounding factors.
Due to its anti-inflammatory properties, IL-10 has also been investigated in GDM and other hyperglycemic conditions[9, 53, 60]. Four studies included in this review reported lower concentration of IL-10 in patients with GDM compared with healthy pregnant women, as expected. However, only one of these four studies reported statistically lower levels of this cytokine. New studies including well-defined selection criteria and a larger number of participants may confirm the hypothesis that reduced IL-10 production is involved in the pathophysiology of GDM.
The small number of studies on all the other six types of cytokines indicates the need for more investigations in this area. Although IL-6 and TNF-A have been recognized as important links between obesity, diabetes and chronic inflammation, other cytokines such as IL-1B, IL-2 and IFN-G have also been implicated in the network of mediators involved in insulin resistance and diabetes.[7, 15] Therefore, new studies are needed to understand the role of these cytokines in the physiopathology of GDM.
There are several factors that influence cytokine production during pregnancy, including gestational age, ethnicity, smoking and BMI. Although several studies matched groups according to BMI and/or to gestational age at sampling,[16-24, 26-29, 42, 43, 46] others did not.[25, 39-41, 44, 45] Smoking habits and ethnicity of the participants were not reported by almost 70% of the studies included in this review. This affected the quality of the results, which was the domain with the worse scores in the quality assessment of the studies. The lack of attention to these factors can in part explain some of the discrepant findings between studies analyzing the same cytokine.
This review had some limitations, such as the exclusion of studies published in languages other than English, Spanish, Portuguese, French or Italian and the lack of search for gray literature (e.g., congress abstracts and unpublished studies). We also acknowledge that by focusing exclusively on human studies that assessed cytokine levels in peripheral blood (serum, plasma and lymphocytes culture supernatant), we limited our evaluation of cytokines in the physiopathology of GDM. Strong points of this review include the use of a broad search strategy, the inclusion of several electronic databases, duplicate study selection, extraction and quality assessment. Finally, to the best of our knowledge, this is the first systematic review of the literature on cytokine levels in GDM.
There is a lack of good quality evidence on possible differences in cytokine levels in women with and without GDM. This reviews points to the need for more adequately designed studies on cytokine levels in patients with GDM compared with healthy controls, especially involving other cytokines besides TNF-A, IL-6 and IL-10. These studies should aim to ensure that variables that can affect cytokine levels, such as gestational age, ethnicity, smoking habits and BMI, are equally distributed in cases and controls or adjusted for in the analyses of the results. Additionally, future studies should also provide detailed information on sample collection, handling, storage and assessment methods, and report their findings according to internationally accepted standards. With the publication of more studies of good methodological quality, future systematic reviews should be able to provide an answer to the existing controversies on cytokine levels in patients with GDM.
The existing studies do not allow any definitive conclusions on differences in cytokine levels in patients with GDM compared with healthy controls. More studies are needed to clarify this question.
This work was financially supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP – 10/52547-5) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).
Conflict of interest
Silvia Daher, Maria Regina Torloni and Bárbara Yasmin Gueuvoghlanian-Silva were authors of one of the studies included in this review.