Association between Dichlorodiphenyldichloroethylene in the Serum and Adipose Tissue with Type 2 Diabetes : A Systematic Review and Meta-Analysis

The number of diabetics is growing every day. In addition to the main risk factors of type 2 diabetes (obesity, age and etc.) other environmental risk factors such as persistent organic pollutants are also considered. Dichlorodiphenyldichloroethylene is one of the persistent organic pollutants which are produced by the pesticide DDT metabolized and its effects of type 2 diabetes are taken into consideration of many investigators. Inconsistency in the results caused to try to achieve a detailed assessment of the effect of dichlorodiphenyldichloroethylene on type 2 diabetes by systematic review and meta-analysis. In this review study after a systematic review, finally 6 prospective and 7 cross-sectional studies were obtained. After approving the heterogeneity of the studies based on random effect model, it was calculate the mean of odds ratio. Meta-analysis of the pooled random effects of Dichlorodiphenyldichloroethylene was done with type 2 diabetes in prospective and cross-sectional studies. In prospective studies, the highest and lowest odds ratio was 6.1 (95%CI: 1.36-27.27) and 1.01 (95%CI: 0.59-1.70) and cross-sectional studies was 5.5 (95%CI: 1.2-25.10) and 0.61 (95%CI: 0.32-1.15), respectively. In the all studies (sum of prospective and cross-sectional), in the random effect model odds ratio was 1.52 (95%CI: 1.26-1.84) and heterogeneity was 52.1% (ρheterogeneity=0.009). Increasing concentration of Dichlorodiphenyldichloroethylene in the serum and adipose tissue increased the risk of type 2 diabetes significantly (P<0.001). The results of this review study support the role of dichlorodiphenyldichloroethylene as an environmental risk factor for type 2 diabetes.


Search and Selection of Studies
A list of titles and abstracts of all articles available in databases such as Pubmed, Scopus, Ovid, Embase and ISI Web of Science from 2015.10.03 to 2015.10.19 was prepared at first by researchers in order to determine and select the related titles to be evaluated independently.Then, the papers related to blinding method of initial evaluation were entered to the process independently.The main inclusion criterion of several studies to this research is a point to the estimation of p,p'-DDE concentration in different age groups, men and women and different concentrations and its association between the T2D.The studies that were not a part of initial researches or in an unrelated field with the title of T2D and review studies were excluded from this study.In the second step using a check list of STROBE (Strengthening the Reporting of observational studies in Epidemiology) which is a standard check list, studies were evaluated.This check list was included 43 various parts and evaluates various aspects of methodology such as measuring methods, measuring variables, statistical analysis and aims of study.In this check list, the minimum and maximum obtained scores were considered as 40 and 45, respectively (Von Elm, Altman, Egger, Pocock, & Gøtzsche, 2007).Finally, the superior studies which had obtained the minimum score of 40 were entered into the research and its related data were extracted for meta-analysis.At the end, the prospective (Cohort) and cross-sectional studies which had studied their association between the concentrations of p,p'-DDE with T2D, were evaluated accurately.There were no restrictions on the language selection because of all the main studies were published English language.Also used to the Boolean term "AND" for combination of the keywords.

Data Extraction
According to the standard method for data extraction, three independent investigators extracted and shape of the table all data.Disagreements points about studies resolved by consensus between three investigators.Information extracted included first author, publication year, study design, concentration of p,p-DDD in serum and adipose tissue, sample size, odds ratio (OR), confidence interval, P value and population study.

Data Analysis and Synthesis
All statistical analysis was performed by Comprehensive Meta-Analysis V.2.2.064 software.Heterogeneity index using the heterogeneity test between the studies was determined by using heterogeneity test.After proving the heterogeneity of studies based on the effect random model, mean to odds ratio was calculated.Meta-analysis of pooled random effects p,p-DDE with T2D was done in the prospective and cross-sectional studies.To search about the heterogeneity of studies, several methods were proposed but one of the most common of them is drawing a funnel plot (Arlington, 2008).This plot used when the number of the evaluated studies is not less than 10 (Durand, Promotion, & Management, 2008).The forest plot was used in the random model for evaluating the significance of serum p,p'-DDE and T2D in each study.Since, the number of studies in review studies are low, the significance level in heterogeneity test is p value<0.001 (Arlington, 2008).

Discussion
Meta-analysis of data showed that p,p-DDE in serum and adipose tissue was significant positive association between the T2D.

Meta-Analysis
In Wu et al. study is as I 2 =36.8% (ρ heterogeneity =0.161) although heterogeneity in Wu et al. study has a less compared to our study, but this heterogeneity was not significant (Wu et al., 2012).Since we try to investigate more and newer studies in our research, it was expected to have differences in our results compared to other review articles.Overall heterogeneity was (I 2 =52%, ρ heterogeneit: 0.009) in the range of 30-60% hence was moderate class heterogeneity (Higgins and Green, 2008).Since in most review studies, the number of evaluated studies was low; so to compensate the low power, significance level for odds ratio and relative risk of p value<0.001 are considered (Schwarzer, Carpenter, & Rücker, 2015).Although against the cross-sectional studies, I 2 was not significant in prospective studies, but generally it can be said that the evaluated studies have a significant and moderate heterogeneity.4).Difference in weight percentage was induced by difference in confidence interval and the difference in confidence interval is due to the difference in type of study, number of population and concentration of p,p'-DDE (Egger, Davey-Smith, & Altman, 2008).Due to the reverse funnel plot in both prospective and cross-sectional studies, publication bias has not been conducted in the all studies (Begg-Mazumdar test, p value=0.31) (Rothstein, Sutton, & Borenstein, 2006).
The studies showed that POPs concentration is decreasing in the environment (Rignell-Hydbom et al., 2009), but it should be noted that firstly, POPs increased the risk of diabetes in obese human by accumulation in adipose tissue (Lee et al., 2006) and secondly, many of old POPs such as PCBs or DDT and the new ones such as brominated flame-retardants, perfluorinated compounds, phthalates and bisphenol a have the nature of endocrine destruction creates a hormone imbalance (Rasier, Toppari, Parent, & Bourguignon, 2006).

Comparison Studies
In Larebeke et al. study against women, the incidence of diabetes in men had a positive significant association between p,p'-DDE concentration (p value<0.001).Although other POPs such as PCB-118 and HCB caused the increase in the incidence of T2D in men and women, but was no significant association between p,p'-DDE concentration and diabetes in women (p>0.52)(Van Larebeke et al., 2015).With increasing the concentration of p,p'-DDE in both studies, the risk of T2D was also increased.
In Turyk et al. (2015), the concentration of Ln-p,p-DD (2 ng/g lipid) was assessed by some diabetes biomarkers such as C-reactive protein (CRP), adiponectin and Gamma-glutamyl transpeptidase (GGT) in fish consumers.
The results showed that risk of diabetes is significantly increased by increasing p,p'-DDE (Table 1).Geometric mean of ln p,p'-DDE concentration in men (2.19 ng/g lipid) was more than women (1.39 ng/g lipid), so higher increase in p,p'-DDE in men was due to more eating fish (11% more) (Turyk et al., 2015).In a study by Azandjeme et al. in the area of Borgou at north of Benin among the residents in linen production area, the concentration of p,p'-DDE in men and women was 6.68 and 6.08 ln ng/g lipid, respectively that women have less concentration than men similar to Turyk et al. study (Azandjeme, Delisle, Fayomi, Ayotte, & Djrolo, 2014).
Higher concentration of p,p'-DDE in men was due to more work of men and their more exposure compared to the women.On one hand, hard work causes more eating and will be followed by more absorption of p,p'-DDE.
In Turyk et al. study (2009) on fish consumers in great sport lake, it was determined that unlike PCB-118 and total PCBs, the concentration of p,p'-DDE had a positive significant association between T2D.The increase in concentration of p,p'-DDE cause an increase in the risk of diabetes (Table 1).As Turyk et al. (2015) and Azandjeme et al. studies in the range concentration of p,p'-DDE of 2.3-5.3 ng/g lipid, the risk of T2D in men (OR=6.8)was more than women (OR=3.9).In the present study, the concentration of p,p-DDE was much lower than other studies and there was a question that how low concentration of p,p'-DDE could create an almost equal risk of diabetes in other studies (Turyk et al., 2009).
Wu et al. study showed that the increase in concentration of p,p'-DDE cannot significantly increase the risk of T2D in both groups of women with Non Hodgkin Lymphoma (P=0.86) and Breast Cancer Study (P=0.98)(Wu et al., 2012) (Wu et al., 2012).Serum concentrations after 7 years from the baseline examination were 5680±6160 (pg/ml) and 3890±3770 (pg/ml) in case and control groups respectively that had a higher difference between case and control compared to the last years (case more than control).Thus, it is expected that risk of T2D increases.
High OR in RignellHydbom et al. study compared to certain studies except fish consumers (Turyk et al., 2009;Philibert et al., 2009) due to the effect of other variables such as obesity, sex and etc.
Codru et al. study showed that against the chemical material of Mirex, the increase in concentration of p,p'-DDE increased the risk of T2D significantly.Also, similar to Gasull et al. study, high HCB increased the risk of diabetes significantly.The concentration mean of ln p,p'-DDE (6.26±6.23 ng/g lipid) was more than Turyk et al. (2015) study and like this study, the risk of T2D was significantly increased (Codru et al., 2007).Langer et al. study showed that in the areas with high concentrations of PCB, p,p'-DDE and HCB, impaired fasting glucose was higher other areas (Langer, Kočan, Tajtaková, Petrík, & Chovancová, 2007).Also Cox et al. study like Lee et al. study showed that concentration of p,p'-DDE in obese human is much more than slim human.On the other hand, concentration of p,p'-DDE has a significant association between T2D.The mean of concentration of p,p'-DDE (28.09 ng/g lipid) in this study was several times more than Codru et al., Turyk et al. (2015) and Azandjeme et al. (Cox, Niskar, Narayan, & Marcus, 2007).
In Philibert et al. study, the mean of concentration of p,p'-DDE was 7.16 ng/g lipid in fish consumers which was more than the other studies and Turyk et al. ( 2009) (fish consumers).There was a significant and positive association between fish consuming and incidence of diabetes, so it has a low weight percentage in forest plot (Figure 3).Since the concentration of p,p'-DDE in this study is high, the risk of T2D in this study is also higher than other studies except Turyk et al. (2009).The odds ratio in Philibert et al. and Turyk et al. 2009 studies is more than other prospective and cross-sectional studies (Figure 3).One of the causes of this difference is fish consumption in the studied groups.Fish can store a lot of POPs in its tissue (Yu, Zhang, Huang, Li, & Pang, 2012).
In Arrebola et al. study, the geometric mean of ln p,p'-DDE concentration in diabetics is 5.49 ng/g lipid.The risk of T2D increased by increasing p,p'-DDE concentration like other studies.It should be noted that in this study, only p,p'-DDE in adipose tissue had a significant association between T2D in the random and fixed models (Arrebola et al., 2013).
Airaksinen et al. study like RignellHydbom et al. study showed that the increase in p,p'-DDE and PCB-153 concentrations cause an in increase in the incidence of T2D, significantly (Airaksinen et al., 2011).Against the studies of Turyk et al. 2015 and Azandjeme et al. in Airaksinen et al. study, p,p'-DDE concentration had no significant difference between men (600±18 ng/g lipid) and women (610±16 ng/g lipid) (p=0.64).Since eating is the main entrance way of POPs to the body (Fernández-González, Yebra-Pimentel, Martínez-Carballo, & Simal-Gándara, 2015), so it can be said that type of food and its amount to use has been the same for men and women in this study.Against the studies of Turyke et al. ( 2009) and Azandjeme et al. in Silverstone et al. study, the risk of T2D due to PCBs and p,p'-DDE in women were more than men.The increased risk of T2D with p,p'-DDE in women and men were significant and insignificant, respectively.The higher risk of diabetes in women is due to more body fat and following that more accumulation of POPs in women's bodies, at first.Secondly, POPs is effective in estrogenic activity of women and following that glucose metabolism (Silverstone et al., 2012).But amount of exposure has the basic role in the increase concentration of p,p'-DDE, as some studies showed that concentration of p,p'-DDE in men has been more than women.
POPs induce diabetes through different mechanisms.The most important of these mechanisms include altered gene transcription, lipid metabolism, insulin production, changes in the insulin signaling pathway (Marchand, Tomkiewicz, Marchandeau, Boitier, & Barouki, 2005) and altered glucose transport (Tonack, Kind, Thompson, Wobus, & Fischer, 2007).It should be noted that each above mentioned mechanisms can be done in men and women.Ukropec et al. study showed that the increase in PCBs and p,p'-DDE increased the risk of T2D significantly.The range of ln p,p'-DDE for p,p'-DDE was 6.08-8.76ng/g lipid.Against the studies of Langer et al., Gasull et al., Codru et al., Wu et al. and Larebeke, in Ukropec et al. study no significant association between HCB concentration and T2D was not observed.This difference in the results due to the lack of consideration of variables such as age, sex and BMI or low concentration of HCB (Ukropec et al., 2010).

Limitations of the Study
There are limitations in this study.First, bias may exist for published data; non-English except Persian language studies were not included.Second, range time research was between 2000 and 2016.

Conclusions
Systematic review and Meta-analysis showed that p,p'-DDE in the serum and adipose tissue increased the risk of T2D significantly (P<0.001).Results of this study, supports the role of p,p'-DDE as an environmental risk factor for T2D.
Figure 1.The results of research in databases of Ovid, Scopus, Embase, ISI Web of Science, Pubmed using the research strategy

Figure 2 .
Figure 2. Forest plot of odds ratio of the T2D in the all studies

Figure 3 .
Figure 3. Forest plot of odds ratio of the T2D in prospective studies

Table 1 .
Characteristics of studies included in the systemic review and meta-analysis similar to Larebeke et al. study that HCB increased the risk of T2D significantly (Van Larebeke et al., 2015).It is while Wu et al. has cited wrong in his article that RignellHydbom et al. has said that only p,p'-DDE increases the risk of T2D al. study, concentration of p,p-DDE has no significant association between diabetes, but PCBs and HCB (like Larebeke et al. study and Wu et al. study) increased the risk of diabetes significantly.It is while the geometric mean concentration of p,p'-DDE was in the range of Turyk et al., but their results were different from each other