Abstract. Lind L, Lind PM (Uppsala
University, Uppsala, Sweden). Can persistent organic pollutants and
plastic-associated chemicals cause cardiovascular disease? (Review). J Intern Med 2012; 271:
537–553.
During
the last decade, associations between persistent organic pollutants (POPs),
such as polychlorinated biphenyls, dioxins and pesticides, and cardiovascular
(CV) risk factors and overt CV disease (CVD) have been reported in humans.
Recently, associations between plastic-associated chemicals (PACs), such as
bisphenol A and phthalates, and CVD have also begun to emerge. Several
approaches to evaluating such associations have been used: accidents with a
high level of exposure, occupational exposure studies, geographical studies of
subjects living near a contaminated area and traditional case–control or cohort
studies with measurements of circulating levels of different environmental
contaminants in the general population. Exposure to POPs has consistently been
associated with diabetes using all the approaches described above, including
prospective studies. The evidence regarding associations between exposure to
POPs and other CV risk factors, such as hypertension, obesity and lipids, is
less strong and is mainly based on cross-sectional data. Associations between
overt CVD and POPs have been reported using all the above approaches, but
prospective data from population-based studies are still lacking to provide
firm evidence of an important and independent role of POP exposure in the
pathogenesis of CVD. Nevertheless, taken together, current evidence suggests
that further longitudinal and experimental studies should be conducted to
investigate the effect of exposure to both POPs and PACs, such as bisphenol A and
phthalates.
Introduction
Since
the Second World War, humans have been exposed to a growing number of man-made
chemicals. To date, about 150 000 chemical substances have been registered
in the database of the European Chemicals Agency (http://www.echa.europa.eu/). Whilst chemical substances
to be used as pharmaceuticals are subjected to rigorous regulatory control,
chemicals intended for most other purposes require little or no regulation
before entering the market. Very few of these chemicals have been tested for
toxicity, and only around 20% of the many chemicals on the market are tested to
a sufficient degree to enable proper assessment of their risks to health and
their effects on the environment.
The
term persistent organic pollutants (POPs) refers to a number of highly
divergent chemicals with the common characteristics of toxicity and resistance
to degradation. Many POPs are highly lipophilic and thereby accumulate in
adipose tissue with a half-life from 1 month up to several years. The
best-known POPs are polychlorinated biphenyls (PCBs), dioxins, brominated flame
retardants and organochlorine (OC) pesticides, such as dichlorodiphenyltrichloroethane
(DDT).
Another
group are high-volume produced chemicals used in the production of plastics and
include bisphenol A and phthalates. Most of these compounds are not lipophilic,
but because of their widespread use in daily life, they are measureable in the
circulation in almost all individuals in the industrialized world. Table 1 provides
a summary of some of the POPs and plastic-associated chemicals (PACs) discussed
in this review.
Table 1. Overview of some of the
toxicants discussed in this review
A
common feature of several POPs and PACs is an effect on hormonal systems, and
therefore, they have been collectively termed ‘endocrine disruptors’. Because
these chemicals can act both as agonists and antagonists on sex hormone
receptors in different tissues, studies have been carried out to investigate
their effects on reproduction and hormone-sensitive cancers. In recent years,
associations with cardiovascular (CV) factors and overt CV disease (CVD) have
also been reported.
In the
present review, we focus on evidence from human studies regarding a link
between POPs/PACs and the development of atherosclerosis and CVD. Five different
research tools have been used: data from accidents in which some subjects were
exposed to very high concentrations of a pollutant, occupational exposure at
moderate to high levels, geographical exposure by living close to a
contaminated area, case–control studies, cross-sectional studies and
traditional longitudinal population studies with measurements of circulating
levels of different chemical compounds (see Table 2 for
summary of study types). Studies with CVD as outcomes are summarized in Table 3.
Table 2. Studies to evaluate the link between
environmental contaminants and human disease
|
Type
of study
|
Comments
|
|||||
|
Accidental exposure (e.g. of a
geographical area or workplace) to very high contaminant levels
|
Resembles classical toxicity
studies in experimental animals. Catches effects on the right part of a
dose–response curve
|
|||||
|
Occupational exposure studies
|
Usually used to evaluate a
moderately high degree of exposure. Could evaluate the effects of a single
compound, but more likely to determine the effects of a mixture of compounds
|
|||||
|
Geographical area exposure studies
(e.g. proximity to a waste site)
|
Usually used to evaluate a low
degree of exposure, and the effects of a mixture of compounds
|
|||||
|
Case–control studies
|
Used to assess exposure either by
history or through blood samples
|
|||||
|
Cross-sectional population-based
studies
|
Usually used to evaluate a low
degree of exposure. Able to assess exposure either by history or through
blood samples
|
|||||
|
Prospective population-based
studies
Table 3. Summary of studies relating POP or
PAC levels to CVD
|
Usually used to evaluate a low
degree of exposure. Able to assess exposure either by history or through
blood samples. Reverse causality has less effect than in case–control and
cross-sectional studies
|
|||||
|
Type
of study
|
Study
[references]
|
Study
design
|
Number
of subjects/cases
|
Endpoint
|
Results
|
|
|
1.
MI, myocardial infarction; CVD,
cardiovascular disease; PAC, plastic-associated chemicals; PAOD, peripheral artery
occlusive disease; POP, persistent organic pollutants; CHD, coronary heart
disease; NS, not significant; M, million; Q, quartile.
|
||||||
|
Accident
|
Prospective/25 years’
follow-up
|
278 108/2402
|
CVD mortality
|
RR 1.8 (95% CI 1.0–3.1) for 5- to
9-year follow-up, NS thereafter
|
||
|
Accident
|
Prospective/24 years’
follow-up
|
748/90
|
Incident CVD
|
RR 1.4 (95% CI 0.7–2.4)
|
||
|
Occupational exposure
|
Prospective/>25 years’
follow-up
|
1499/243
|
Incident CHD
|
RR 1.4 (95% CI 1.0–1.8) for
sprayers of TCDD
|
||
|
Occupational exposure
|
Prospective
|
21 863/158–1183
|
CHD/stroke mortality
|
RR 1.6 (95% CI 1.2–2.2) for CHD,
RR 1.5 (95% CI 0.8–2.8) for stroke
|
||
|
Geographical study
|
Cross-sectional
|
>5 M
|
Prevalent MI
|
RR 1.2 (95% CI 1.0–1.3) for POP
contamination
|
||
|
Geographical study
|
Cross-sectional
|
>5 M/28 216
|
Prevalent stroke
|
RR 1.1 (95% CI 1.0–1.2) for POP
contamination
|
||
|
Population-based study
|
Cross-sectional
|
1455/42–79
|
Prevalent CVD/MI
|
For 1 SD of bisphenol A: RR
1.3 (95% CI 1.1–1.6) for CVD, RR 1.4 (95% CI 1.1–1.7) for MI
|
||
|
Population-based study
|
Cross-sectional
|
1331/38–76
|
Prevalent CVD/MI
|
For 1 SD of bisphenol A: RR
1.2 (95% CI 0.9–1.5) for CVD, RR 1.3 (95% CI 1.0–1.6) for MI
|
||
|
Population-based study
|
Cross-sectional
|
2032/143
|
Prevalent PAOD
|
RR 1.2 (95% CI 1.0–1.5) in obese
subjects for sum of all OC pesticides. NS in nonobese subjects
|
||
|
Population-based study
|
Cross-sectional
|
889/108
|
Prevalent CVD
|
RR for Q4 versus Q1: dioxin-like
PCBs 5.0 (95% CI 1.2–20), nondioxinline PCBs 3.8 (1.1–12), OC pesticides 4.0
(1.0–17); NS for women alone, PCDD and PCDFs
|
||
|
Population-based study
|
Cross-sectional
|
1016/616
|
Carotid artery plaque
|
RR 1.8 (95% CI 1.3–2.6) for a
summary measure of PCBs for Q4 versus Q1 RR 1.49 (95% CI 1.0–2.0) for sum of
all pesticides
|
||
|
Population-based study
|
Cross-sectional
|
1016/616
|
Carotid artery plaque
|
RR 1.4 (95% CI 1.0–1.9) for Q4
versus Q1 for mono-methyl-phthalate levels
|
||
Risk factors for CVD
Diabetes
Accidents. In Seveso in northern Italy, a
manufacturing plant exploded in 1976, and the surrounding area was heavily
contaminated with the highly toxic dioxin 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). After
25 years of follow-up, the diabetes mortality rate in Seveso was elevated
[relative risk (RR) 1.7, 95% confidence interval (CI) 1.1– 2.7] compared with
the rate in a nonexposed surrounding reference area.
In the
late 1970s, the consumption of rice-bran oil laced with PCBs and
polychlorinated dibenzofurans (PCDFs) poisoned thousands of Taiwanese people in
Yucheng. At follow-up 24 years later, 378 subjects from Yucheng were
compared with 370 matched nonexposed controls. The risk of diabetes was
significantly increased for women [odds ratio (OR) 2.1, 95% CI 1.1–4.5], but
not for men (OR 1.0), amongst the Yucheng population. Furthermore, women from
Yucheng diagnosed with chloracne, a sign of severe PCB/dioxin exposure, had an
adjusted OR for diabetes development of 5.5].
Occupational studies. During the Vietnam War, the US
Air Force used the herbicide Agent Orange, a chemical contaminated with TCDD. A
large number of studies have been conducted with various follow-up periods,
most often using Air Force soldiers who served in Vietnam but were not exposed
to Agent Orange as controls. In general, these studies showed an increased risk
of diabetes development in the exposed Vietnam veterans (RR values between 2.0
and 4.0) as significant in some, but not all, studies.
Geographical studies. A study team led by Carpenter
investigated >900 waste sites in New York, NY, USA, and quantified the
extent of POP contamination. Merging data of individual’s zip codes and health
records, the investigators concluded that subjects living close to a
contaminated waste site showed a 23% (95% CI 15–32%) increased risk of
diabetes.
Population-based studies. Although a large number of
cross-sectional studies investigating POP levels in different populations have
been reported, we were only able to identify five prospective studies. In the
Michigan polybrominated biphenyl (PBB) cohort of >1300 subjects living in an
area of known POP contamination, 180 subjects received a diagnosis of diabetes
during a 25-year follow-up. Comparing the highest versus the lowest quintile of
a summary measurement of PBBs revealed an OR of 2.0–3.0 in women and 1.7 in
men; only the OR value in women was statistically significant. As there are
many chemical forms of PBBs and PCBs, it is common to present a summary measure
instead of analysing individual compounds. The potential drawback of this
procedure is that different compounds might have different actions or effects
because of variations in half-lives.
In a
cohort of fish consumers from sport fishing in the Great Lakes followed for
more than 10 years, 36 new diabetes cases were observed. Circulating
levels of 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene, DDE (a metabolite of
DDT; OR 7.1, 95% CI 1.6–31 for the highest tertile), but not a summary measure
of PCBs (OR 1.8, 95% CI 0.6–5.0 for the highest tertile), predicted development
of diabetes.
In the
population-based CARDIA study, a nested case–control study was performed after
17 years of follow-up and 90 new cases of diabetes were identified.
Circulating levels of the pesticide trans-nonachlordane (TNC) and several PCB
congeners predicted diabetes development. This effect was evident even for very
low doses of TNC (OR 5.3 for second versus first sextile) and was most
pronounced in obese subjects.
In a
nested case–control study conducted in women from Lund, Sweden, DDE (OR 5.5,
95% CI 1.2–25 for the highest quartile), but not PCB153 (OR 1.6, 95% CI
0.6–4.0), predicted development of diabetes that was diagnosed more than
7 years after the baseline examination].
In the
Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study,
a cohort study including >1000 subjects all aged 70 years, we
investigated the associations between POPs and diabetes using both a
cross-sectional design (113 cases) and prospective data during 5 years of
follow-up (36 cases). Using both approaches, a number of PCB congeners, as well
as a summary measure of pesticides including DDE and TNC, were found to be
associated with diabetes]. Of note, the summary measure of PCBs was a stronger
risk factor for incident diabetes (OR 7.5, 95% CI 1.4–38) than body mass index
(BMI) (OR 4.5, 95% CI 1.2–22), even following adjustment for BMI.
The
results of most, but not all, of the cross-sectional studies confirmed the
findings of the prospective studies, suggesting a role of POPs in diabetes.. It
is noteworthy that all cross-sectional studies analysing TNC found a
significant association with diabetes.
In the
US National Health and Nutrition Examination Survey (NHANES) of 2003–2004,
urinary bisphenol A levels were analysed in >1400 subjects. Using
self-reported diabetes as outcome in a cross-sectional analysis, there was an
RR of 1.3 (95% CI 1.2–1.6) for a one standard deviation (SD) change in urinary
bisphenol A levels]. However, an investigation of the NHANES 2005–2006 survey
cycle did not confirm these findings (OR 1.0, 95% CI 0.8–1.3).
Brominated
flame retardants, such as polybrominated diphenyl ethers (PBDEs) or PBBs, were
also evaluated in a cross-sectional analysis of the NHANES data. Prevalent
diabetes was significantly associated with serum concentrations of PBB-153 and
PBDE-153 (ORs 1.8–3.1 for highest versus lowest quartiles).
Hypertension
Compared
with studies of diabetes, investigations of the association between
hypertension and POPs are rare.
In the
Yucheng PCB accident, women diagnosed with chloracne showed an adjusted OR of
3.5 for hypertension compared with those without chloracne during 24 years
of follow-up.
In one
of the above-mentioned studies of US Vietnam veterans, an increased incidence
rate of hypertension was found amongst those who had sprayed Agent Orange
compared with controls (increased risk of 32%, 95% CI 8–61%).
In the
above-mentioned study of contaminated waste sites in New York, a 19% increased
risk (95% CI 9–31%) in hypertension was observed amongst those living close to
a contaminated site].
Population-based studies. In a cross-sectional analysis
of the NHANES 1999–2002 data, it was found that the RR for hypertension was 1.8
(95% CI 1.2–2.7) for the highest quartile of a number of PCBs, whilst another
evaluation of the same survey including 524 subjects showed that dioxin and
PCDF concentrations in serum were related to newly diagnosed hypertension in
women only (RR 5–6 for highest versus lowest quartiles).
In a
cross-sectional analysis of data from 758 participants living in a polluted
environment in Anniston, AL, USA, RR for hypertension for the highest versus
the lowest tertile of PCBs was 4.1 (95% CI 1.3–14).
In some
of these studies, the diagnosis of hypertension was based on the combination of
blood pressure measurements and/or history of antihypertensive medication,
which is appropriate for determining the prevalence of hypertension, as more
than half of all hypertensive cases are unknown in population surveys in
adults.
Metabolic
syndrome
Population-based studies. With regard to the metabolic
syndrome (MetS), there are only cross-sectional data on the effect of these
pollutants. Most investigations have used the National Cholesterol Education
Program/Adult Treatment Panel III, NCEP/ATP III criteria for the diagnosis of
the MetS].
In a
nationwide survey of >1300 Japanese subjects, circulating levels of dioxins
and PCBs were related to the occurrence of the MetS (OR 3.2–4.8 for the highest
versus lowest quartiles). Levels of these POPs were associated with all
components of the syndrome.
Using
data from >700 nondiabetic subjects in the NHANES 1999–2002 survey, it was
demonstrated that levels of pesticides were strongly related to the MetS (OR
5.3, 95% CI 2.5–11.0, for the highest versus lowest quartile). Pesticide levels
were significantly related to all criteria of the syndrome except blood
pressure [ORs (highest versus lowest quartile): 2.4 for waist circumference,
7.1 for triglycerides, 2.3 for HDL, 5.6 for glucose and 1.8 for blood
pressure].
In a
case–control study of 50 nondiabetic Korean subjects with the MetS and 50
controls, two POP pesticides, beta-hexachlorocyclohexane and heptachlor
epoxide, were associated with the MetS (OR 4.4 and 6.0, respectively, for the
highest quartiles).
Because
a common feature of the MetS is insulin resistance, the relationship between
POPs and insulin resistance has also been analysed using NHANES data in 749
nondiabetic participants. The homoeostasis model assessment-estimated insulin
resistance (HOMA-IR) index was used to define insulin resistance. Nineteen POPs
were investigated. Amongst the POPs, pesticides were most strongly associated
with HOMA-IR (OR 3.8 for highest quartile). The association between pesticides
and HOMA-IR tended to strengthen as waist circumference increased.
Obesity
As with
the MetS, most of the data regarding associations between obesity and
environmental contaminants are from cross-sectional studies.
Population-based studies. In a cross-sectional study of
the PIVUS cohort, we investigated the associations between POP levels and fat
mass using dual-emission X-ray absorptiometry. PCBs with a low degree of
chlorination were found to be positively related to fat mass, whereas PCBs with
a high degree of chlorination were inversely related to fat mass. Levels of
pesticides were therefore positively related to fat mass. This unexpected
finding of divergent effects of different PCBs might be explained by different
pharmacokinetic properties, and the fact that PCBs with a low degree of
chlorination have a shorter half-life than highly chlorinated PCBs, as well that
peak exposure to PCBs was in the 1970s. As noted by Wolff et al., the
relationship between POP concentrations and fat mass (BMI) is always negative
immediately after the main exposure because of the fact that POPs are mainly
stored in fat tissue. However, following 2–3 half-lives of the compound, this
relationship becomes positive if there is no further exposure. The timing of
the change from a negative to a positive relationship is governed mainly by the
half-life of the compound, but also by changes in body fat and by any recurrent
exposure. An inverse relationship between highly chlorinated PCBs and obesity
was also reported by Dirinck et al. in a smaller case–control study of obese
and lean subjects.
In a
cross-sectional analysis of NHANES data, gender interactions were seen in the
relations between pesticide levels and BMI, whilst a positive relationship
between levels of a dioxin and BMI was seen in both men and women.
In
another cross-sectional Japanese study of >13 000 subjects in the
community, a positive relationship was found between the level of all PCBs and
BMI. Similar findings were seen in a smaller substudy using the CARDIA cohort,
where subjects with high levels of PCBs had a high BMI and waist circumference
after 25 years of follow-up].
Using
NHANES 2003–2004 data, positive associations were found between different
phthalate metabolites and BMI or waist circumference. Most often, these
associations were found only in either men or women.
One
problem with the study of associations between POPs and obesity/MetS is that
POPs are stored in adipose tissue and thus the amount of fat determines the
circulating levels for a given exposure, as discussed above. Thus, it is very
hard in cross-sectional studies to determine whether high POP exposure could
induce obesity/the MetS, as has been suggested based on animal studies.
Prospective studies would in part overcome this difficulty, but very few such
studies have been reported.
In a
prospective analysis of the PIVUS cohort, 100 new cases of abdominal obesity
were discovered over a 5-year follow-up period. Similar to our previously
reported findings of the cross-sectional analysis, PCBs with a low degree of
chlorination were positively related to the risk of future abdominal obesity,
whereas PCBs with a high degree of chlorination were inversely related to the
risk of future abdominal obesity. Levels of pesticides were positively related
to the risk of future abdominal obesity.
Inspired
by the Barker hypothesis that intra-foetal programming is a major determinant
of obesity, hypertension and CVD in adult life, POP levels have been measured
in mothers during pregnancy or in cord blood at delivery, and the children have
been followed (mother–child cohorts). As recently reviewed in detail], no
consistent association has been found between foetal exposure to PCBs and
obesity during childhood or early adult life, whilst a majority of the studies
in which foetal exposure to DDE was measured reported positive relationships
with later BMI or body weight]. It should be noted, however, that only one of
these studies evaluated obesity during a prolonged follow-up
(20–50 years), so the full potential of these mother–child studies remains
unclear at present.
Lipids
Most
POPs are highly lipid-soluble compounds that are transported by lipoproteins,
and therefore, any associations between levels of POPs and lipids could be hard
to distinguish in a meaningful way. Furthermore, traditionally POP levels have
been measured as concentration normalized with respect to serum lipids.
Occupational and geographical
studies. Although
traditional population-based studies might not be suitable to determine the
relationship between POP levels and lipids, some occupational exposure and
geographical studies have addressed this issue.
In a case
study of workers so heavily exposed to TCDD, it was shown that even
40 years after exposure, they still showed levels >100% higher than
normally found in humans, 91% of the subjects were on anti-lipid medication and
100% showed atherosclerotic plaques in the carotid arteries at ultrasound].
Although no control group was included in this case study of 11 Caucasian
subjects with a mean age of 66 years, both the use of anti-lipid
medication and the prevalence of carotid atherosclerosis were higher than usually
found in this age group (13% and 66%, respectively, in the PIVUS study in
70-year-old subjects).
In a
cross-sectional study amongst 133 male workers at municipal waste incineration
sites, serum cholesterol levels were higher in those with a high Toxic
Equivalent (TEQ; a marker for dioxin exposure) compared with those with a low
value].
A total
of 281 workers from two chemical plants in the USA and 260 control referents
participated in a case–control study; those with high levels of dioxins showed
increased levels of serum triglycerides].
In a
cross-sectional study of >1000 subjects in a POP-polluted area of East
Slovakia, serum triglyceride levels were increased compared with a control
population from a nonpolluted nearby area.
Population-based studies. In contrast to other POPs,
perfluorinated compounds are not highly lipid soluble (they are oil and water
repellants) and lipid normalization is not needed. An association between
perfluorooctane sulphonate/perfluorooctanoic acid (PFOA) levels and serum cholesterol
was found in a large US cross-sectional study]. In a cross-sectional study of
1025 workers with potential exposure to PFOA, the circulating levels of this
compound were related to LDL cholesterol concentration.
Atherosclerosis
Population-based
studies
Very
few studies on the relationship between human atherosclerosis and levels of
POPs or PACs have been reported. In the PIVUS study, the carotid arteries were
examined by ultrasound. Following adjustment for lipids and a further 10 known
CVD risk factors, a number of PCBs were found to be related to the occurrence
of plaques in a cross-sectional analysis. In addition, levels of highly
chlorinated PCBs were related to an echolucent (dark) vascular wall indicative
of lipid infiltration in the intima space. Furthermore, there was a
relationship between the phthalate metabolite mono-methyl phthalate and plaque
occurrence, even at levels only slightly higher than the lowest levels in the
sample (low-dose effect). Levels of bisphenol A and several phthalate
metabolites were related to lipid infiltration of the vascular wall.
CVD
Nonspecific
CVD
Accidents and occupational
studies. Mortality
rates were investigated in a 25-year follow-up of the Seveso accident. It was
found that CVD mortality was increased during the first 10 years following
the accident, but thereafter no further increase was observed. The highest CVD
mortality rates were found in the 5- to 9-year follow-up interval (RR 1.8, 95%
CI 1.0–3.1).
In the
24-year follow-up of the Yucheng accident, a nonsignificant trend for
self-reported CVD was reported in the PCB- and PCDF-exposed cases, compared
with controls (OR 1.5, 95% CI 0.8–2.7).
In one
of the above-mentioned US Vietnam veteran studies, an increased incidence rate
of CVD was found after 30 years of follow-up in those who had sprayed
Agent Orange compared with veterans who did not handle the herbicide (OR 1.5,
95% CI 1.1–1.9).
Population-based studies. In one cross-sectional analysis
of 889 adults from the NHANES 1999–2002, concentrations of PCBs and pesticides
in serum were positively related to self-reported CVD, defined as a history of
myocardial infarction, angina pectoris or stroke. PCBs and OC pesticides were
associated with the prevalence of CVD in women only (OR 3.8–5.0 for the highest
versus the lowest quartiles), whereas the total level of all three evaluated
dioxins was related to CVD in both sexes (OR 1.7 for the highest versus the
lowest quartile)].
In a
cross-sectional analysis of 1455 adults from the NHANES 2003–2004, urinary
levels of bisphenol A were positively associated with self-reported CVD, with
an RR of 1.6 (95% CI 1.1–2.2) for a change of 1 SD in urinary bisphenol A
levels. An analysis of the NHANES 2005–2006 confirmed these findings]. In these
studies, no distinction was made between different CV disorders.
Coronary
heart disease
Because
the results of several small occupational studies pointed towards an
association between dioxin exposure and coronary heart disease, the IARC
international cohort was formed, consisting of 36 cohorts from 13 countries,
including 21 863 workers followed for >20 years. A significant
association was found between dioxin exposure and coronary heart disease in
this powerful analysis (RR 1.6, 95% CI 1.2–2.2).
In the
above-mentioned study of contaminated waste sites in New York, an increased
risk of 20% (95% CI 3–39%) for a diagnosis of myocardial infarction was found
for those living close to a contaminated site].
In the
cross-sectional analysis of the combined NHANES 2003–2004 and 2005–2006 data,
there was an RR of 1.2 (95% CI 1.1–1.4) for a change of 1 SD in urinary
bisphenol A levels for prevalent myocardial infarction].
Stroke
In the
IARC international cohort including 21 863 workers exposed to dioxin
followed for >20 years, a nonsignificant tendency for an association
with incidence of stroke was noted (RR 1.5, 95% CI 0.8–2.8)].
Similar
to the findings for myocardial infarction, living close to a POP-contaminated
waste site in New York was associated with an increased risk of being
hospitalized for stroke (RR 1.1, 95% CI, 1.0–1.2)].
Peripheral
artery disease
In a
recent evaluation of 2032 subjects from the NHANES 1999–2004, it was shown that
OC pesticides were associated with peripheral artery disease (PAD), defined as
an ankle–brachial index <0.9 (a commonly used definition of PAD). This
association was most pronounced in obese subjects (OR 1.1, 95% CI, 1.0–1.3).
Mechanisms
of action
As
discussed above, many of the POPs have been shown to interact with hormonal
systems. Indeed, different effects on sex hormone receptors have been
demonstrated with both agonist and antagonist properties of different POPs in
different tissues and at different levels of exposure. Bisphenol A was termed
an oestrogen as early as the 1930s, when it was shown to stimulate the
reproductive system in female rats. There were plans for its introduction as an
oestrogenic pharmaceutical agent, although it never reached the market because
more potent agonists were found]. A number of other steroid nuclear receptors
are also known to be affected by POPs. With regard to dioxin and dioxin-like
PCBs, a main target is the aryl hydrocarbon receptor (AHR). Following activation
by dioxin, this receptor, with no clearly identified endogenous ligand or known
action, will increase the activity of the cytochrome P450 enzyme CYP1A1,
resulting in formation of reactive oxygen species (ROS) and inflammation]. In
addition, alterations in the apoptosis rate and cell cycling have been
described following dioxin activation of the AHR. Thus, several mechanisms that
are currently discussed in relation to the pathogenesis of atherosclerosis are
known to be triggered by POPs, including induction of ROS during oxidation of
LDL cholesterol in the vascular wall, ROS-induced low-grade chronic
inflammation within the atherosclerotic plaque, and apoptosis of vascular
smooth muscle cells making the plaque vulnerable to rupture and thrombosis
formation.
Other
effects of environmental contaminants might be due to the activation of other
nuclear receptors. One such example is the well-established action of
phthalates on Peroxisome proliferator-activated receptors, PPAR receptors,
which will affect adipocyte differentiation, lipid metabolism and insulin
resistance, and thereby the risk of diabetes development].
Validation
issues
As
shown above, there is evidence to suggest that exposure to POPs, and to some
extent PACs, is associated with major CV risk factors, as well as with the risk
of atherosclerosis and overt CVD. However, compared with other traditional risk
factors, such as high LDL cholesterol, smoking and hypertension, it is not
possible to perform clinical trials, as for example, for lipid-lowering or
smoking cessation medication, to establish whether these environmental
contaminants are true risk factors or merely risk markers. It would be
unethical to administer these contaminants in clinical trials, and there are no
known ways to induce their rapid elimination. Furthermore, the Mendelian
randomization approach, as recently used to determine whether C-reactive
protein is a risk factor or a risk marker, is not applicable to nonendogenous
potential risk factors, as in this case. Therefore, at present, the only other
way to validate the human findings reported in the present review is to
challenge appropriate experimental in vivo and in vitro models with dosages relevant
for human exposure to seek biological evidence of causal relationships.
In summary,
there is experimental evidence to support the human findings, showing that POPs
and/or PACs (i) induce diabetes because of impairments in insulin secretion
and/or resistance], (ii) induce obesity by means of altered adipocyte
differentiation], (iii) raise blood pressure by increasing the sensitivity to
angiotensin II] and impairing endothelium-dependent vasodilatation], (iv)
induce dyslipidaemia both by interfering with lipid storage in adipocytes and
by altering lipid-regulating enzymes in the liver, leading to hepatic
steatosis], (v) accelerate atherosclerosis formation in ApoE knockout mice (a
well-established atherosclerosis model) , (vi) induce cardiac hypertrophy and
(vii) alter the composition of contractile myocardial elements and induce cardiac
fibrosis. However, it is noteworthy that in many of these studies, no attempts
have been made to transfer relevant human exposure data to the experimental
setting, and the exposure levels are often much higher than are seen in humans.
Problems
associated with in human studies
The two
major problems with conducting proper population-based large prospective
studies in this field are the high cost and the analytical volume needed.
Therefore, it is necessary to improve the analytical procedures to increase the
speed of the analysis and to reduce the sample volume without loss of
precision. In the PIVUS study, van Bavel and co-workers have been able to
analyse >20 POPs with high precision in 0.5 mL plasma. However, even more improvements in analytical
capacity at lower volumes must be achieved before it is realistic to use frozen
samples from the major long-term prospective studies of CVD that have been
performed in the USA and Europe. If these improvements can be achieved,
knowledge of the relationships between environmental contaminants and CVD will
increase significantly.
Another
problem concerns the pharmacokinetics of POPs. It is not possible to determine
true exposure to POPs from one sample alone because of their very long
half-lives; maximal exposure in the Western world occurred about four decades
ago, but widespread use of these compounds continues in Africa, Asia and Latin
America with global spread by sea and air. Thus, repeated samples over time
together with pharmacokinetic modelling may overcome this problem and thus
obtain a proper estimate of exposure for POPs. Some PACs, including bisphenol
A, are less persistent in the body than POPs. Nevertheless, repeated sampling
might also result in a more precise estimate of exposure to PACs.
Humans
are not exposed to one chemical at a time, but are chronically exposed to a
simultaneous cocktail of chemicals. The potential for synergistic or additive
effects from such multiple exposures must be taken into account to evaluate the
combined effects of contaminant exposure on health. This issue has recently
been highlighted by the World Health Organization , the Swedish Chemical Agency
,the US Centers for Disease Control and Prevention, the European Commission and
in a report by Silins and Hogberg. Thus, it is essential to develop new
statistical strategies to identify highly potent mixtures with additive or
synergistic effects of relevance for CVD development. Traditional statistical
models of interactions would demand much larger samples than are available
today to search for interactions between multiple contaminants.
It has been
shown both experimentally and in human studies that a nonmonotonic relationship
between POP exposure and different outcomes, such as conditions that predispose
towards CVD (CV risk factors) or overt CVD, might exist. In these studies,
effects were also seen at very low levels of exposure, and no further effect
size was noted if the level of exposure was increased. In some cases, the
effect even disappeared during high exposure. This feature of some POPs will
make the interpretation of dose–response relationships very challenging,
especially when the effects of mixtures are evaluated. shows the odds ratios for a
PCB (congener 138) divided into quintiles versus three outcomes in the PIVUS
study. With regard to prevalent carotid artery atherosclerosis, the main effect
was seen in the highest quintile; however, the highest OR values were not seen
in the highest quintile for the other two outcomes, incident abdominal obesity
and incident diabetes. Indeed, a considerable decrease in risk was seen in the
highest quintile for these two outcomes, exemplifying the nonmonotonic
relationships often observed in studies of POPs.
Summary showing
published data of the relationships between a PCB (congener 138) divided into
quintiles and three different outcomes in the PIVUS study: incident diabetes
incident abdominal obesity and prevalent carotid artery atherosclerosis Odds
ratio values are given for quintiles 2–5 versus quintile 1 (given as 1.0)
following adjustment for age, gender, lipids, exercise habits, education level,
smoking, body mass index (for atherosclerosis and diabetes), hypertension and
glucose (for atherosclerosis only). Please see text for discussion of the
nonmonotonic relationships for obesity and diabetes.
Another
problem in this field is that large prospective studies with long follow-up
periods are ideal to generate confidence in the results. However, if the length
of the follow-up period is increased, the contaminants measured at baseline may
have been banned by the authorities by the time of the analysis and therefore
evaluation would no longer be relevant. These chemicals might have been
replaced by others during the follow-up period. Thus, there is always a risk
that when waiting for the results of ‘ideal’ studies to emerge, the range of
contaminants will have changed. An example is that PCBs were banned in the
1970s and were in part replaced by brominated compounds (i.e. bromine rather
than chlorine in PCBs). The chlorinated and brominated chemicals have now to
some extent been replaced by fluorinated compounds, but neither the brominated
nor the fluorinated compounds were measured at baseline in the prospective
studies that are currently being conducted with prolonged follow-up periods.
Repeated sampling in longitudinal studies might be one way to deal with this
problem.
A major
source of the environmental contaminants discussed in this review is food.
Fatty fish, such as salmon and herring, are heavily contaminated with POPs in
some parts of the world. Fish are also known to be contaminated with mercury,
for example, in the Baltic Sea. Thus, the associations between POPs and some of
the phenotypes discussed here might be confounded by the health effects of fish
oils, as well as the effects of mercury. These other additional effects of diet
and contaminants that are closely associated with POP exposure are
unfortunately very hard to control for accurately, because detailed dietary
assessments together with measurements of multiple contaminants are rare in
environmental research.
Strengths
and limitations
The
major strength of the reviewed data is the level of congruence of the different
approaches (accidents, occupational exposure, geographical exposure and
population-based studies). In the case of the existing data regarding diabetes,
prospective population-based studies with measured circulating levels
contribute the highest degree of evidence. Five such prospective studies have
been reported, and all but one show significant associations between PCB
exposure and future diabetes. In addition, there is agreement amongst the five
studies regarding DDT exposure and diabetes. These data are largely supported
by cross-sectional studies, accidents and occupational and geographical
exposure studies.
Few
prospective studies of the effects of POP exposure on conditions that increase
the risk of CVD, atherosclerosis and overt CVD have been conducted, and
therefore, any reported associations are based on a lower level of evidence.
However, cross-sectional data suggest that further longitudinal studies of the
effects of POPS would be highly worthwhile. Similarly, despite the availability
of fewer reported studies, further longitudinal studies of PACs, including
bisphenol A and phthalates, should be considered.
No comments:
Post a Comment