Chinese Medical Journal 2008;121(8):701-705
Variations of tumor necrosis factor-α, leptin and adiponectin in mid-trimester of gestational diabetes mellitus

GAO Xue-lian,  YANG Hui-xia,  ZHAO Yi

GAO Xue-lian (Department of Obstetrics & Gynecology, Peking University First Hospital, Beijing 100034, China)

YANG Hui-xia (Department of Obstetrics & Gynecology, Peking University First Hospital, Beijing 100034, China)

ZHAO Yi (Department of Obstetrics & Gynecology, Peking University First Hospital, Beijing 100034, China)

Correspondence to:YANG Hui-xia,Department of Obstetrics & Gynecology, Peking University First Hospital, Beijing 100034, China (Tel: . Fax:.
diabetes mellitus, gestational; adiponectin; leptin; tumor necrosis factor-α

Background  Many cytokines have been found to increase the insulin resistance during pregnancy complicated by glucose metabolism disorder. This study aimed to investigate which comes first, the changes of some cytokines or the abnormal glucose metabolism.
Methods  This nested case-control study was undertaken from January 2004 to March 2005. Twenty-two women with gestational diabetes mellitus (GDM), 10 with gestational impaired glucose tolerance (GIGT), and 20 healthy pregnant women were chosen from the women who had visited the antenatal clinics and had blood samples prospectively taken and kept during their visit. The levels of tumor necrosis factor-α (TNF-α), leptin and adiponectin were determined. One-way ANOVA analysis and bivariate correlation analysis were used to assess the laboratory results and their relationship with body mass index (BMI).
Results  Women with GDM have the highest values of TNF-α and leptin and the lowest value of adiponectin compared with those with GIGT and the healthy controls (P <0.01) at 14–20 weeks of gestation. This was also found when these women progressed to 24–32 weeks. The significantly increased levels of TNF-α and leptin and the decreased level of adiponectin were found at the different periods of gestation within the same group. Positive correlation was shown between the levels of TNF-α and leptin at the two periods of gestation with the BMI at 14–20 weeks, while adiponectin was negatively correlated (P <0.05).
Conclusions  The concentrations of TNF-α, leptin and adiponectin may change before the appearance of the abnormal glucose level during pregnancy. Further studies are required to verify the mechanism of this alteration and whether the three cytokines can be predictors for GDM at an early stage of pregnancy.

Glucose metabolism disorders in pregnancy, including gestational diabetes mellitus (GDM) and gestational impaired glucose tolerance (GIGT), are a common complication during pregnancy and one of the main causes of adverse fetal-maternal outcomes.1 However, the pathogenesis of GDM has not been clearly stated. The most common theory is that GDM is caused by reduced insulin sensitivity combined with the anti-insulin hormones increasingly secreted by the placenta during pregnancy, such as human placental lactogen (HPL), prolactin (PRL), glucocorticoid, progesterone, etc.2,3

Cytokines including tumor necrosis factor-α (TNF-α), leptin and adiponectin were found to be able to increase the insulin resistance (IR) in the tissues and decrease the insulin sensitivity, thus leading to GDM.4 Studies focused on the relationship between these factors and gestational glucose metabolism disorders or in other diseases both in animals and human beings.5-7 However, few studies concentrated on the significance of these factors in predicting glucose metabolism disorders with the progress of pregnancy.4 Earlier screening and interventions for GDM in high-risk women might improve the maternal and fetal outcomes. Hence we hypothesized that if the levels of these cytokines increase in early or mid-trimester, these factors can be used as the predictors of glucose metabolism disorders. Therefore, a nested case-control study was conducted to assess the values and variations of TNF-α, leptin and adiponectin before the onset of glucose metabolism disorders as well as the factors related with the levels of these cytokines.


Selection of cases and diagnostic criteria
A nested case-control study was conducted at the Depart- ment of Obstetrics & Gynecology, Peking University First Hospital, Beijing, China. Since January 2004, serum samples (1 ml) were collected at 14–20 weeks of gestation from all pregnant women who consecutively attended to the prenatal clinics for early screening. These women were followed up until delivery. Another serum samples were also collected from all the pregnant women at the time they were subjected to glucose challenging test (GCT) or oral glucose tolerance test (OGTT) between 24–32 weeks of gestation. The samples were stored at –80°C. Twenty-two women with GDM and 10 with GIGT who delivered from September 2004 to March 2005 served as the study group, and blood samples at earlier pregnancy and 24–32 weeks were available after exclusion of those women with incomplete clinical data. Twenty healthy pregnant women who delivered at the same period and attended the prenatal clinics for GCT or OGTT with normal OGTT results were chosen as the control. For these 52 pregnant women, serum samples taken at 14–20 and 24–32 weeks of gestation were retrieved and the levels of TNF-α, leptin and adiponectin were measured.

All subjects met the following criteria: no history of pre-gestational diabetes; normal fasting glucose level during the first trimester; no history of liver, renal, thyroid or other diseases which might influence the metabolism of protein and lipoprotein. Collection of blood samples was done after informed consent was obtained from the pregnant women.

Diagnostic criteria for abnormal glucose metabolism during pregnancy are as follows: for all women with 50 g GCT ≥7.8 mmol/L, 75g OGTT was performed and the cut-off value of fasting, 1, 2, and 3 hours glucose level after glucose intake were 5.8, 10.5, 9.2 and 8.0 mmol/L, respectively. GDM was diagnosed in those women who had two abnormal values out of the four, and GIGT was diagnosed when only one abnormal glucose level was detected among the four results.8,9

After GDM or GIGT was diagnosed, insulin injection was given if glucose level had not been well managed after 3 days of dietary control. However, no women with GIGT required insulin in this study.

Laboratory test
The concentrations of TNF-α, leptin and adiponectin were tested by ELISA. The commercial human TNF-α ELISA kit with the sensitivity of 25 pg/ml was obtained from Diaclone Company, French. The range of measurement was 25–800 pg/ml.

The commercial human leptin ELISA kit and human adiponectin ELISA kit both with the sensitivity of 100 pg/ml were purchased from Yuetai Company, China. And the range of measurement was 312.5–10 000 pg/ml and 100–6000 pg/ml, respectively. All steps followed the instructions enclosed in the kit.

Statistical analysis
All data were analyzed with the SSPS 11.0 software. The basic clinical data of all subjects and the serum levels of TNF-α, leptin and adiponectin were recorded as mean±standard deviation (SD) and One-way ANOVA analysis was used. The relationship between body mass index (BMI) and the serum levels of TNF-α, leptin and adiponectin was analyzed by bivariate correlation analysis. P <0.05 was considered statistically significant.


General information
No significant difference was found in the parity, BMI in mid-term and gestations at sampling among the GDM, GIGT and control groups (Table 1). However, significant difference was observed in the age among the three groups (P <0.05).

view in a new window

Table 1. Clinical information of the selected pregnant women

Concentrations of TNF-α, leptin and adiponectin in early and mid-trimester
Significant difference was found in the values of TNF-α, leptin and adiponectin among different groups both at 14–20 and 24–32 weeks of gestation (P <0.05, Table 2). Within each group, an increase of the levels of TNF-α and leptin and a decrease of the level of adiponectin were observed with the progress of pregnancy (Table 2).

view in a new window

Table 2. Comparison of concentrations of TNF-α, leptin and adiponectin among the three groups at 14–20 and 24–32 weeks

Correlations between BMI and cytokines at different gestations
The serum levels of TNF-α and leptin at 14–20 and 24–32 weeks were positively correlated with BMI at 14–20 weeks (TNF-α: r=0.271, 0.291, P <0.01; leptin: r=0.247, 0.338, P <0.01) and those of liponection were negatively correlated with BMI (r=–0.215, –0.293, P <0.01).

Relationship between the levels of cytokines and glucose
The results of 50 g GCT were positively correlated with the levels of TNF-α and leptin (r=0.727, r=0.666, P <0.001), whereas a negative correlation was shown between the adiponectin level and the GCT results (r=–0.505, P <0.001).

Again, the levels of TNF-α and leptin were also positively correlated with the fasting, 1 and 2 hours glucose results of OGTT, but not with the 3-hour glucose level. The levels of adiponectin were negatively correlated with the glucose levels of fasting and 1 and 2 hours shown by OGTT (Table 3).

view in a new window

Table 3. Relationship between the levels of TNF-α, leptin, and adiponectin with fasting and 1- and 2-hour glucose results after OGTT (r value)


Glucose metabolism disorder is a common complication during pregnancy and its pathogenesis is associated with IR and deficiency of insulin secretion.10,11 It has been presumed for a long time that glucocorticoid and placenta secreted hormones (HPL, progesterone, estrogen, prolactin, etc) mediate the IR. However, recent studies have focused on several new factors including TNF-α, leptin and adiponectin,4,12,13 which could lead to IR. Kirwan et al4 stated that among all the markers being tested including TNF-α, leptin, glucocorticoid, estrogen, progesterone and HPL, TNF-α was an independent predictive factor for insulin sensitivity, but no statistical significance was noted among insulin sensitivity, placental hormones and glucocorticoid in the late trimester.14 This result challenged the traditional theory that reproductive hormones are essential to reduced insulin sensitivity during pregnancy.

Most studies measured the levels of the above factors only after GDM had been clearly diagnosed.15-18 Coughlan et al19 found that the body reaction to glucose increased when complicated with GDM and the placenta and fat tissues were stimulated to produce more TNF-α by the increased glucose level. Also, researchers in China reported that the serum concentrations of TNF-α and adiponectin varied during pregnancy.20,21 But all reports failed to explain whether the higher plasma glucose level leads to the increased level of TNF-α or the increased TNF-α level reduces the insulin sensitivity and then results in GDM.

However, in our study, the nested case-control method was used to explore the relationship between the levels of the factors and the onset of glucose metabolism disorder, thus the pregnant women were selected early in 14–20 weeks. At this early period of gestation with a normal glucose level, we presumed that it is the variation of cytokines which made the IR worse if abnormal concentration of the cytokines was detected. In order to find the changes of the several cytokines during pregnancy, all women were tested repeatedly at 24–32 weeks. They had significant differences in the levels of TNF-α, leptin among the GDM, GIGT and control groups at 14–20 weeks, with the highest level in the GDM group. Increased levels of TNF-α and leptin were also shown when abnormal glucose metabolism was diagnosed later. Thus, we can verify that cytokines changes happened earlier than the increase of IR during pregnancy. Kirwan et al4 reported the similar results. They tested the fasting serum concentrations of TNF-α, leptin, corticoid, other placental hormones, and IR in 5 GDM and 10 normal controls with the euglycemia-glucagon clamping technique at pre-gestation, 12–14 weeks and 34–36 weeks, respectively. They found that TNF-α other than placental hormones and corticoid was associated with IR. Hence the variation of TNF-α as a marker in predicting insulin sensitivity plays an important role in the pathogenesis of GDM.

Qiu et al12 studied 823 pregnant women and found a significant correlation between the concentration of leptin and the risk of GDM (P <0.001) by measuring the leptin level at 13 weeks of gestation. Those women with a leptin level ≥31.0 ng/ml at 13 weeks carried 4.7 more times the risk of GDM later than those with a leptin level ≤14.3 ng/ml. This was equal to that every increase of 10 ng/ml in the leptin level means a 20% increase in the risk of GDM. Our study showed a higher leptin level in GDM women than in the controls and the GIGT women. This finding was consistent with that reported elsewhere, but as the limited number of patients, no definite cutoff value was determined.

The associations between leptin or TNF-α with prepregnant diabetic or GDM mothers and their children have been widely investigated.22-25 But studies on adiponectin in this area are rare.

Adiponectin, an adipose tissue-derived plasma protein, is involved in regulation of insulin resistance and glucose hemostasis, and thus is a key modulator of insulin action and glucose metabolism.26 It is found to be related with not only diabetes, but also obesity, coronary heart disease, obstructive sleep syndrome apnea27 and breast cancer.28 However, the variation of adiponectin level during pregnancy is just contrary to the level of TNF-α and leptin, and the level of adiponectin is lower in GDM women than in controls, showing a decreasing trend with the progress of pregnancy. This suggests that the low level of adiponectin may induce severe IR before the onset of GDM and that adiponectin may be related to the pathogenesis of abnormal glucose metabolism during pregnancy as well.

Williams et al13 found in 968 pregnant women the same result that the adiponectin level was <6.4 mg/ml in 73% of GDM women and 33% of the controls. The risk of GDM in women with an adiponectin level of <6.4 mg/ml was 4.6 times as that in those with the level of >6.4 mg/ml. This finding was consistent with the fact that low level of adiponectin is the risk factor for T2DM. However, more studies are required to postulate this theory for GDM. Another study also showed higher levels of adiponectin in T1DM throughout gestations compared to nondiabetic pregnancies with no difference in leptin levels.29

Bivariate correlation analysis in this study found a positive correlation between BMI at early gestation and the levels of TNF-α and leptin, but a negative correlation between BMI and the level of adiponectin. Another study30 also showed that the levels of TNF-α, TNF-α soluable receptor, and C-peptide as well as the ratio of C-peptide to plasma glucose level were positively correlated to BMI which influenced the serum level of TNF-α.

Adamczadk and coworkers31 allocated 34 primiparas at the gestational age of 27-33 weeks to two different groups: the healthy pregnant women and those complicated with diabetes. They found a higher serum level of leptin in the diabetes group and also the serum level of leptin was positively correlated with the 1-hour plasma glucose level in 50 g GCT. Aside from the same findings in their study, we also found that the levels of TNF-α and leptin were also positively correlated with fasting and 1- and 2-hour glucose levels after OGTT, but no relation with the 3-hour glucose level was observed in OGTT. Again, the adiponectin level showed a negative correlation with fasting and 1- and 2-hour glucose levels after OGTT. This demonstrated that leptin and adiponectin were also associated factors for GDM and their serum levels were correlated with the severity of GDM. Obviously they might have antagonistic action against insulin.

In conclusion, this study confirmed that the increased levels of TNF-α and leptin and the decreased level of adiponectin preceded the onset of abnormal glucose level during pregnancy. Thus these cytokines might be of predictive value in GDM diagnosis. However, further studies are required to investigate the mechanism of the alteration of the three cytokines.


1. Sun WJ, Yang HX. Maternal and fetal outcomes in pregnant women with abnormal glucose metabolism. Chin J Obstet Gynecol (Chin) 2007; 42: 377-381.

2. Cunningham FG, Gant NF, Leveno KJ, GilstrapIII LC, Hauth JC, Wenstrom KD. Williams Obstetrics. 23th ed. New York: McGraw-Hill; 2005: 127-128.

3. Hod M, Jovanovic L, Di Renzo GC. Textbook of diabetes and pregnancy. London: Martin Dunitz; 2003: 32-64.

4. Kirwan JP, Hauguel-De Mouzon S, Lepercq J, Challier JC, Huston-Presley L, Friedman JE, et al. TNF-α is a predictor of insulin resistence in human pregnancy. Diabetes 2002; 51: 2207-2231.

5. Clausen T, Burski TK, Oyen N, Godang K, Bollerslev J, Henriksen T. Maternal anthropometric and metabolic factors in the first half of pregnancy and risk of neonatal macrosomia in term pregnancies. A prospective study. Eur J Endocrinol 2005; 153: 887-894.

6. Ategbo JM, Grissa O, Yessoufou A, Hichami A, Dramane KL, Moutairou K, et al. Modulation of adipokines and cytokines in gestational diabetes and macrosomia. J Clin Endocrinol Metab 2006; 91: 4137-4143.

7. Lu L, Zhang Q, Pu LJ, Xu XW, Zhang JS, Hu J, et al. Elevation of tumor necrosis factor-alpha, interleukin-1 beta and interleukin-6 levels in aortic intima of Chinese Guizhou minipigs with streptozotocin-induced diabetes. Chin Med J 2007; 120: 479-484.

8. Yang HX, Gao XL, Dong Y, Shi CY. Analysis of oral glucose tolerance test in pregnant women with abnormal glucose metabolism. Chin Med J 2005; 118: 995-999.

9. Li K, Yang HX. Value of fructosamine measurement in pregnant women with abnormal glucose tolerance. Chin Med J 2006; 119: 1861-1865.

10. Craigo SD, Baker ER. Medical complications in pregnancy. New York: McGraw-Hill; 2005: 41.

11. Reece EA, Coustan DR, Gabbe SG. Diabetes in women. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2004: 363-370.

12. Qiu C, Williams MA, Vadachkoria S, Frederick IO, Luthy DA. Increased maternal plasma leptin in early pregnancy and risk of gestational diabetes mellitus. Obstet Gynecol 2004; 103: 519-525.

13. Williams MA, Qiu C, Muy-Rivera M, Vadachkoria S, Song T, Luthy DA. Plasma adiponectin concentrations in early pregnancy and subsequent risk of gestational diabetes mellitus. J Clin Endocrinol Metab 2004; 89: 2306-2611.

14. Ryan EA, Enns L. Role of gestational hormones in the induction of insulin resistance. J Clin Endocrinol Metab 1998; 67: 341-347.

15. Wang SL, Liu PQ, Ding Yu, Peng W, Qu X. Study on tumor necrosis factor-α and insulin resistance in pregnant women with gestational diabetes mellitus. Chin J Obstet Gynecol (Chin) 2004; 39: 737-740.

16. Yang S. Relationship between TNF-α and insulin resistance in gestational diabetes. Chin J Immunol (Chin) 2004; 20: 652-653.

17. Liu T, Shang T, Tian Z. Maternal and fetal serum leptin level and placental leptin expression in gestational diabetes mellitus. Chin J Perinatol (Chin) 2004; 7: 214-216.

18. Niu JM, Guo XL, Zhang W. Relationship between the serum level of adiponectin and insulin secretion and resistance in women with gestational diabetes mellitus. Chin J Pract Gynecol Obstet (Chin) 2004; 20: 615-617.

19. Coughlan MT, Oliva K, Georgiou HM, Permezel JM, Rice GE. Glucose-induced release of tumour necrosis factor-α from human placental and adipose tissues in gestational diabetes mellitus. Diabet Med 2001; 18: 921-927.

20. Li HH, Tang P, Ding SJ, Wang ZY. The study on the relationship between the serum adiponectin and TNF-α and insulin resistance in the GDM patients. Maternal Child Health (Chin) 2007; 22: 921-923.

21. Li LQ, Luo YP, Hu CL, Zhou C. Analysis of TNF-α value in women with GDM. J Prac Diabetol (Chin) 2006; 2: 17-18.

22. Lindsay RS, Hamilton BA, Calder AA, Johnstone FD, Walker JD. Scottish multicentre study of diabetes in pregnancy. The relation of insulin, leptin and IGF-1 to birthweight in offspring of women with type 1 diabetes. Clin Endocrinol 2004; 61:353-359.

23. Torchinsky A, Gongadze M, Orenstein H, Savion S, Fein A, Toder V. TNF-alpha acts to prevent occurrence of malformed fetuses in diabetic mice. Diabetologia 2004; 47:132-139.

24. Manderson JG, Patterson CC, Hadden DR, Traub AI, Leslie H, McCance DR. Lepting concentrations in maternal serum and cord blood in diabetic and nondiabetic pregnancy. Am J Obstet Gynecol 2003; 188: 1326-1332.

25. Challier J, Galtier M, Bintein T, Cortez A, Lepercq J, Hauguel-de Mouzon S. Placental leptin receptor isoforms in normal and pathological pregnancies. Placenta 2003; 24: 92-99.

26. Mazaki-Tovi S, Kanety H, Sivan E. Adiponectin and human pregnancy. Curr Diab Rep 2005; 5: 278-281.

27. Zhang XL, Yin KS, Li C, Jia EZ, Li YQ, Gao ZF. Effect of continuous positive airway pressure treatment on serum adiponectin level and mean arterial pressure in male patients with obstructive sleep apnea syndrome. Chin Med J 2007; 120: 1477-1481.

28. Hou WK, Xu YX, Yu T, Zhang L, Zhang WW, Fu CL, et al. Adipocytokines and breast cancer risk. Chin Med J 2007; 120: 1592-1596.

29. Randeva HS, Vatish M, Tan BK, Qureshi Z, Boovalingam P, Lewandowski KC, et al. Raised plasma adiponectin levels in type 1 diabetic pregnancies. Clin Endocrinol 2006; 65: 17-21.

30. Winkler G, Cseh K, Baranyi E, Melczer Z, Speer G, Hajos P, et al. Tumor necrosis factor system in insulin resistance in gestational diabetes. Diabetes Res Clin Pract 2002; 56: 93-99.

31. Adamczadk M, Kokof F, Wlasma AW. Relationship between plasma rennin profile and leptinemia in patients with essential hypertention. J Hum Hypertens 2000; 14: 503-509.

  1. National “211” Project Peking University Evidence-Based Medicine Group,No. 91000-242156028;