Persistent organic pollutants-environmental risk factors for diabetes mellitus? - A population-based study.

BACKGROUND: Globally, type-2 diabetes mellitus is increasing in epidemic proportions. A major cause of concern in India is the increasing incidence of cases, especially troubling is the observed increase in younger age groups with no risk factors. New evidence suggests that many environmental factors, such as air pollution, persistent organic pollutants (POPs), and environmental estrogens are implicated as risk factors for type-2 diabetes mellitus. Animal and human epidemiological studies have shown ubiquitous lipophilic substances, including POPs, are frequently associated with type-2 diabetes mellitus. Such studies have not been undertaken in Indian youth.
METHODS: This is a cross-sectional study that explored the association between POPs and type-2 diabetes mellitus in Indian urban and rural population. About 7 ml of venous blood was collected from all consenting patients and serum was separated immediately and was transported to the lab for further analysis. Serum levels of POPs, including organochlorine (OC) compounds and organophosphorus pesticides, were estimated using sample gas chromatography-mass spectrometry (GC-MS). The fasting blood sugar values and the serum levels of POPS were tested using Pearson correlation coefficient. The magnitude of increase in blood sugar corresponding to increase in POPs was analyzed using linear regression analysis. The odds ratios (ORs) were expressed at 95% confidence intervals (CIs).
RESULTS: Three OC pesticides and one organophosphate pesticide were strongly associated with increasing blood sugar levels after adjusting for age, sex, and body mass index - lindane (OR 4.95, 95% CI 1.03-23.73), DDT o, p' (OR 3.50, 95% CI 1.04-11.73), dimethoate (OR 19.31, 95% CI 4.22-88.37), and dichlorvas (OR 6.33, 95% CI 1.28-31.18). Copyright:

INTRODUCTION

Incidence of diabetes is rapidly increasing worldwide.[1] The dominant risk factors include obesity, sedentary lifestyle, poor diet, old age, ethnicity, family history of diabetes, and dyslipidemia. Prevalence of diabetes for all age groups worldwide was estimated to be 2.8% in 2000 and expected to rise to 4.4% by 2030. The total number of people with diabetes is projected to rise from 171 million in 2000 to 366 million in 2030.[1] Type-2 diabetes is an epidemic in Asia, characterized by a rapid increase in onset at relatively young ages and low body mass index (BMI),[2] and environmental estrogens implicated in the rise of the same.[3] Persistent organic pollutants are a class of compounds characterized by their ability to persist in the environment, due to their low water, high lipid solubility and bio-magnification in the food chain.[4] POPs include polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans, polychlorinated biphenyls (PCB), hexachlorobenzene, p, p'-dichlorodiphenyldichloroethylene (DDE), and several organochlorine (OC) mixtures. These have been studied by international organizations for their impact on environment and humans alike and have been implicated in several adverse effects on human health especially with impaired glucose metabolism and prevalence of diabetes mellitus.[45] India, a highly populated middle-income country, is subjected to environmental contamination of POPs from several activities and sources like production units, illegal imports as well as stockpiles of obsolete pesticide.[6] Except for DDT seven other pesticide POPs listed in the Stockholm Convention are banned for manufacture but are in wide use because large stockpiles are still available. In a nationwide food survey in 2001, 75% of the samples had detectable levels of DDT with about 10–15% of the samples having levels of DDT that exceeded the levels prescribed by the Food and Agricultural organization and World Health Organization. Aldrin and dieldrin were also frequently detected pesticides in foodstuffs.[78] Racial and ethnic differences in responses to POPs are well known.[9] POPs may be associated with T2DM with a higher magnitude in Asians, contributing to the epidemic. Hence, prevention and control of diabetes is a top public health priority in Asian populations.[10]

Many studies based on self-reported diabetes and mortality rate analysis indicate an association between pollutants and diabetes/prediabetes.[1112] Many POPs act as endocrine disrupting agents in animals and humans.[13] Concentrations of most of these pollutants have been diminishing in the environment, food chain, and human body over recent decades in Europe and the United States. However, there are subgroups within these populations that still show an elevated body burden because of dietary habits and current or past exposures.[11] Diabetes has been associated with dioxin-like chemicals, nondioxin like PCBs, DDE, and/or other OCs in several cross-sectional investigations.[4111415] Altered glucose transport, lipid metabolism, and modifications in the insulin signaling pathways are potential mechanisms that might be involved in the association between POPs and diabetes.[11] The Jørgensen group showed that POPs may affect insulin secretion rather than insulin resistance.[16]

POPs metabolism, particularly low-level exposures, warrants investigations using a cohort study design.[17] There have been few studies that have used fasting blood glucose levels (≥126 mg/dL or 7.0 mmol/L), HbA1c, or random glucose level (≥200 mg/dL or 11.1 mmol/L) to diagnose diabetes.[414] The Morgan lab at University of Iowa College of Medicine, Iowa observed that workers with incident diabetes had significantly elevated levels of DDE+DDT in their serum compared to workers without diabetes.[18] Vasiliu et al.[12] found that PCBs were associated with incident diabetes in women but not in men and remained associated when cases encountered during the first 15 years of follow-up were excluded, suggesting that reverse causality was an unlikely explanation for the relationship. Investigations on Operation Ranch Hand Veterans, who were exposed to 2,378- tetra chlorodibenzo-p-dioxin (TCDD) through application of Agent Orange in Vietnam, were found to have incident diabetes associated with TCDD exposure.[19] Also, diabetes was associated with DDE exposure but not with mono-ortho PCB-118, total PCBs, or years of sport fish consumption.[5] In another case-control study it was confirmed that p, p'- DDE exposure could be a potential risk factor for T2DM.[11] A cross-sectional survey conducted in a heavily polluted area of eastern Slovakia demonstrated increasing serum concentrations of individual POPs considerably increased prevalence of prediabetes and diabetes in a dose-dependent manner.[20] Based on the current literature, there is a strong implied linkage between POPs exposure in humans and type-2 diabetes. There have been many studies analyzing the presence of POPs in Indian population, but none connecting the link and establishing association between POPs and diabetes.[21] Hence in this study, we investigated the association of blood sugar with the serum concentrations of POPs in south Indian population.

METHODS

This is a cross-sectional study conducted at the rural and urban outreach centers affiliated to the PSG Institute of Medical Sciences and Research, Coimbatore, India. The study was conducted with approval from Institutional Human Ethics Committee at PSGIMS&R. A written informed consent was obtained from all eligible participants. All participants who consented were recruited in the study. Details about the study were informed by health workers from the center by house to house visit in addition to distribution of pamphlets which included details of the study and requirement of 8 h fasting for the collection of samples. Both male and female of age above 18 were included based on convenience sampling. Diabetics were confirmed as evidenced by fasting blood sugar above 126 mg/dL or HbA1c above 6.4% or use of hypoglycemic medications. Participants with type-1 diabetes and terminally ill conditions were excluded from the analysis. Basic demographic details, socioeconomic status, family history of diabetes, hypertension, general health, and medical history were elicited using a questionnaire. Anthropometric measurements were recorded and blood pressure was recorded in the sitting posture in the right arm. Age, gender, body weight, and height were recorded for each patient. Body mass index was calculated as weight in kilograms divided by height in meter squared. Blood was drawn after at least 8 h overnight fasting. Fasting blood sugar (FBS) was determined by hexokinase method which is a standard laboratory procedure. Serum POPs were determined by method described by Koc and Karakus.[22] Briefly, OC pesticide residues were determined using florisil packed-columns and GC-MS. Linear response between FBS and POPs were measured by the Pearson correlation method. The association between increasing sugar level and POPs was calculated by linear regression after adjusting for age, sex and BMI. All statistical analyses were conducted using R version 3.5.1 (R Development Core Team, http://www.r-project.org).

RESULTS

The total numbers of participants in our study was 191, out of which FBS was not available for five participants. The demographic characteristics for the remaining 186 participants are given in Table 1. Box plots for age and BMI stratified as per gender are given in Figure 1a and 1b. Of the 44 compounds, 11 of them contributed to 93% of total POPs load [Figure 2]. The serum levels of the following four compounds – lindane, dichlorvas, DDT o p', and dimethoate were significantly correlated with increasing blood sugar levels [Table 2]. Among the remaining compounds 31 were not comparable and seven were below limit of detection. Correlation statistics among selected compounds is illustrated in Figure 3, nonsignificant correlation indicated by cross mark, positive correlation by blue solid circles, and negative correlation by red solid circles. Size the circle indicates the magnitude of correlation. Figure 4a–d illustrates the trend analysis by Pearson method showing increasing linear relationship as sugar level increases. Since POPs are obesogenic, we checked its correlation between different strata of BMI [Table 3] and with age [Table 4]. The participants were categorized as normal, overweight, and obese based on the height and weight. The odds ratio (OR) and 95% confidence intervals (CIs) of POPs showed significant risks for compounds like lindane (OR 4.95, 95% CI 1.03–23.73), DDT o, p' (OR 3.50, 95% CI 1.04–11.73), dimethoate (OR 19.31, 95% CI 4.22–88.37), dichlorvas (OR 6.33, 95% CI 1.28–31.18). Similar compounds are found to be significantly associated with blood sugar after adjusting for age, sex, and BMI [Table 5].

Table 1

Demographics of participants enrolled in the study

Demographic characteristics of the participants (n=186)
Gender (n)	Male	56	Female	130
	Min	Mean	25%	Median	75%	Max
Age (year)	32	58.46	50.25	60	67	86
Height (cm)	137	154.3	149.2	152.5	158	184
Weight (kg)	31	61.75	54	61	68	126
BMI (kg/m2)	16.44	25.86	22.35	25.43	28.95	46.28

n: number of participants

Figure 1

Box plots for (a) Age and (b) BMI categorized as per gender

Figure 2

Concentrations of major POPs compounds in the serum of the participants. End.Ketone - Endrin Ketone; Par.Methyl - Parathion methyl; Mono - Monocrotophos; Hept.Epoxide – Heptachlor Epoxide

Table 2

Correlation between POPs (ng/mL) and serum blood sugar levels

Compound	Min	Mean	Median	Max	SD	r 2
Lindane	1.02	10.92	4.50	38.57	9.26	0.22 *
Dichlorvas	1.05	17.98	18.40	47.92	8.54	0.19 *
DDT o, p’	1.05	17.89	20.13	42.76	10.25	0.16 *
Dimethoate	15.98	27.17	25.35	75.68	9.44	0.36

*P<0.05; SD – Standard deviation; r2 – Pearson correlation coefficien*

Figure 3

Correlogram for POPs. Non-significant correlation indicated by cross mark, positive correlation by blue solid circles and negative correlation by red solid circles

Figure 4

Trend plots for compounds showing significant correlation (a) Lindane (b) Dichlorvas (c) DDT (d) Dimethoate. (FBS – Fasting Blood Sugar)

Table 3

Correlation between POPs (ng/mL) and BMI

Compound	Mean	Median	SD	r 2
				Normal	Overweight	Obese
Dichrotophos	7.50	7.22	7.30	0.70 *	NA	0.42
Monochrotophos	71.22	52.68	78.63	0.07	-0.07	0.19 *
1,1 Biphenyl, 3,3’,4 Trichloro	1.34	1.13	0.41	-0.67 *	NA	-0.34
1,1 Biphenyl, 2,2’,4,4’,5,5’ Hexachloro	1.47	1.37	0.29	NA	0.9 *	NA

*P<0.05; SD – Standard deviation; r2 – Pearson correlation coefficient

Table 4

Correlation between POPs (ng/mL) and age (years)

Compound	Min	Mean	Median	Max	SD	r 2
BHC alpha isomer	1.02	9.5	11.78	24.5	5.89	- 0.4*
Parathion methyl	3.32	142.7	28.51	2324.91	314.7	- 0.1*
DDE, p, p’	1.50	36.59	10.99	484.98	71.15	- 0.1*
DDD, o, p’	1.11	43.92	16.77	1160.8	124.67	- 0.2*
1,1 Biphenyl, 2,4,4’ Trichloro	1.06	6.60	5.10	21.35	5.29	0.1*

*P<0.05; SD – Standard deviation; r2 – Pearson correlation coefficient

Table 5

Risks of increasing sugar levels and POPs

Compound	Parameters	Odds ratio	95% CI		P
			Lower	Upper
Lindane	POPs	4.9585	1.0357	23.7392	0.0491 *
	Age	1.5382	0.4401	5.3757	0.5023
	BMI	0.8169	0.028	23.804	0.9068
	Gender	NA	NA	NA	NA
DDT, o, p’-	POPs	3.5041	1.046	11.7382	0.044 *
	Age	2.3531	0.8923	6.2056	0.086
	BMI	0.3416	0.0236	4.9413	0.4321
	Gender	NA	NA	NA	NA
Dimethoate	POPs	19.3124	4.2205	88.3709	2.00E -04*
	Age	2.1495	0.6246	7.3971	0.228
	BMI	0.5344	0.0211	13.5535	0.7049
	Gender	NA	NA	NA	NA
Dichlorvas	POPs	6.3379	1.2882	31.1826	0.025 *
	Age	3.2142	1.0804	9.562	0.0381 *
	BMI	1.0741	0.0507	22.7355	0.9635
	Gender	NA	NA	NA	NA

*P<0.05; CI – confidence interval

DISCUSSION

Our study demonstrated that environmental exposure to various OC and organophosphate pesticides are associated with prevalent type-2 diabetes in Indian population. Our findings are supported with published literature, which implicate OC pesticides in risk of diabetes among general population.[1415] Benzene Hexachloride (BHC) beta isomer was shown to be a strongly correlated with age in our study. One possible mechanistic pathway for BHC isomers is via pancreatic and extra-pancreatic actions, i.e. caspase 3 activation due to reactive oxygen species generation causing membrane destabilization and mitochondrial dysfunction of the beta cell.[15202324] Other OCs like DDD and DDT were also moderately associated with sugar levels, their ORs a little above 1.00. Epidemiological evidence also substantiates the positive risk DDT and its metabolites can have on diabetes.[2024] DDT and its metabolites are inducers of cytochrome P450 (CYP 2BA and CYP 4A1), enzymes involved in lipid and steroid metabolism.[2526] Moreover, these compounds can inhibit aromatase, an enzyme that metabolizes steroids like cortisol.[27] As a result of aromatase inhibition, steroidal compounds tend to accumulate leading to anti-estrogenic effects, although at a receptor level, they are pro-estrogenic in action.[28] These effects coupled with cross talk with catecholamines in the adrenal gland lead to endocrine disruption. These compounds are also known to inhibit expression of Dio-2, low-density lipoprotein (LDL), Protein kinase B (AKT), extracellular-signal-regulated kinase (ERK), Patatin-like phospholipase domain-containing protein (PNPLA), and Glut-4 genes in target tissues.[29] Although we could not find epidemiological evidence for dimethoate even after PubMed search with the following terms “Dimethoate” [MESH] AND “Diabetes” [MESH], we were able to find animal studies implicating omethoate (a break-down product of dimethoate) with insulin resistance.[30] The epidemiological evidences are summarized in Table 6. Major limitation of the study is the cross-sectional nature, which might not have captured the cause-effect relationship.[3132]

Table 6

Epidemiological and animal studies substantiating our study

Compounds	Odds ratio (CI)	Exposure contrast	Reference
p, p’- DDT	2.52 (1.26-5.02)	20.8-26.6 vs. ≤20.7 (ng/g lipid adj.)	Everett and Matheson[24]
	1.96 (1.29-2.98)	≥20.7 vs. ≤20.7 (ng/g lipid adj.)	Everett and Matheson[24]
	1.84 (1.03-2.27)	QU3 vs. QU1 (ng/g lipid adj.)	Ukropec et al.[20]
p, p’- DDD	3.6 (0.8-16.3)	T3 vs. T1 (ng/g lipid std.)	Son et al.[31]
o, p’- DDT	12.3 (1.3-113.2)	T3 vs. T1 (ng/g lipid std.)	Son et al.[31]
β- HCH	2.3 (1-4.3)	≥ 1 vs.<1 ng/g ww	Cox et al.[15]
	2.67 (1.59-4.49)	≥ 9.35-< 9.35 ng/g lipid adj.	Everett and Matheson[24]
	1.08 (0.59-1.97)	QU3 vs. QU1 (ng/g lipid adj.)	Ukropec et al.[20]
	0.8 (0.3-2.2)	Q4 vs. Q1 pg/g	Lee et al.[4]
	8.2 (1.3-53.4)	T3 vs. T1 ng/g lipid adj.	Son et al.[31]
Lindane	2.02 (0.88-4.65)	Not available	Zhang et al.[30]
Dichlorvas	1.21 (0.98-1.49)	Not available	Montgomery et al.[32]
Dimethoate	Study suggested Omethoate could potentially cause insulin resistance.		Zhang et al.[30]

CONCLUSION

This exploratory study found that serum levels of several OCs and organophosphate pesticides were associated with type-2 diabetes in a south Indian population. This is the first study in the Indian population to report the association of POPs and type-2 diabetes mellitus. Our finding complement other studies on Asian groups providing evidence for adverse effects of POPs, which explains the current epidemic of type-2 diabetes in Asia.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.

Financial support and sponsorship

The study was funded by “Collaborative grant between ICMR-University of Minnesota, USA.”

Conflicts of interest

There are no conflicts of interest.