4. Few individual water sources, so
few measurements needed
Why Do Arsenic Studies In Northern Chile?
Region II, Chile:
“The driest habitable place on earth”
Other Areas
Worldwide:
Most water from small private wells
Thousands of water sources, so its hard
to assess the everyone’s exposure
5. Region City or Town Population
Average Arsenic Concentration (µg/L)
Years
1930-57 1958-70 1971-77 1978-79 1980-87 1988-94 1995+
I Arica 168,594 10 10 10 10 10 10 9
Putre 1,799 1 1 1 1 1 1 1
Iquique 196,941 60 60 60 60 60 60 10
Huara 2,365 30 30 30 30 30 30 30
Pica 5,622 10 10 10 10 10 10 10
Pozo Almonte 9,855 40 40 40 40 40 40 40
II Tocopilla 21,827 250 250 636 110 110 40 10
Maria Elena 6,852 250 250 636 110 110 39 39
Calama 125,946 150 150 287 110 110 40 38
San Pedro 4,522 600 600 600 600 600 600 600
Antofagasta 270,184 90 860 110 110 70 40 10
Mejillones 7,660 90 860 110 110 70 37 10
Taltal 10,101 60 60 60 60 60 60 60
Recent migrants 82,312 <10 <10 <10 <10 <10 <10 <10
Population data are based on the most recent Chile census
Concentrations of Arsenic in Drinking Water (µg/L) in
Northern Chile by Year
6. Highly exposed in
utero and early life
40+ YEARS
Long term impacts of early-life exposure: Antofagasta
Bornhere
Cancer?
7. Smith et al., 1998 Am J Epidemiol 147:660-9
ECOLOGIC STUDIES OF CANCER MORTALITY
IN REGION II VS. REST OF CHILE
8. ECOLOGIC STUDIES: Criticisms
Ecologic bias: no exposure information at the individual level
Confounding by…
Smoking
Occupation
Diet
Other
In and out migration
9. CASE-CONTROL STUDY: METHODS
Study design: Case-control study
Study area: Regions 1 and 2 in Northern Chile
Outcome: Lung (n=302), bladder (n=232), and kidney cancer (n=123)
Time frame: All incident cases from 2007-2010
Case ascertainment: All pathology and radiology departments in the Regions
Control ascertainment: Chile Electoral Registry (frequency matched on age and
sex)
Data collection: In home interviews, structured study questionnaire
Exposure assessment: Lifetime residential history linked to an arsenic concentration
in the public water supply of each city in Chile (90% coverage)
Other data: Occupations, smoking, second hand smoke, medical history, diet, blood,
sputum, rna-later, nails, urine
Funding: NIH/NIEHS (R01 ES014032-01 and P42 ES04705)
Ethics approval: UCB, PUC, Region II (Informed consent obtained)
10. 22. ¿En qué lugares ha vivido Ud. A LO MENOS 6 MESES?
Período Lugares Fuente de agua potable en ese lugar
Comuna
Compañía de
Agua municipal
Embotellada Otra:
Desde Hasta Localidad % % %
a. 19 _____ 19 _____
b. 19 _____ 19 _____
c. 19 _____ 19 _____
d. 19 _____ 19 _____
Subject: John Doe
Year City As (ug/L) % bottled Adjusted
Birth 1967 Antofagasta 860 0 860
1968 Antofagasta 860 0 860
1969 Antofagasta 860 0 860
1970 Antofagasta 860 0 860
1971 Antofagasta 860 0 860
1972 Antofagasta 100 0 100
1973 Antofagasta 100 0 100
1974 Arica 8 0 8
1975 Arica 8 0 8
1976 Arica 8 0 8
1977 Arica 8 0 8
1978 Arica 8 0 8
1979 Arica 8 0 8
1980 Arica 8 0 8
1981 Arica 8 0 8
1982 Arica 8 0 8
1983 Arica 8 0 8
1984 Santiago 3 0 3
1985 Santiago 3 0 3
1986 Santiago 3 0 3
1987 Santiago 3 0 3
1988 Santiago 3 0 3
1989 Santiago 3 0 3
1990 Santiago 3 0 3
1991 Santiago 3 0 3
1992 Santiago 3 0 3
1993 Santiago 3 0 3
1994 Santiago 3 0 3
1995 Santiago 3 0 3
1996 Santiago 3 0 3
1997 Santiago 3 0 3
1998 Santiago 3 0 3
1999 Santiago 3 0 3
2000 Santiago 3 0 3
2001 Iquique 60 50 30
2002 Iquique 60 50 30
2003 Iquique 10 0 10
2004 Iquique 10 0 10
2005 Iquique 10 0 10
2006 Iquique 10 0 10
2007 Iquique 10 0 10
2008 Iquique 10 0 10
2009 Iquique 10 0 10
Interview 2010 Iquique 10 0 10
Arsenic water concentrations
Residential history
11. UC Berkeley
Arsenic Health
Effects Research
Group
UC Berkeley
Superfund
Research
Program
US
Collaborators
Allan Smith
Dave Kalman
(University of
Washington)
Lee Moore
(NCI)
John Balmes
(UCSF)
Ken Cantor
(NCI)
Meera Smith
Christine Skibola
Luoping Zhang
Chile PUC,
Region II
Catterina Ferreccio
Guillermo Marshall
Yan Yuan
Jane Liaw
Martyn Smith
Johanna Acevedo
Sandra Cortes
Many others
Craig Steinmaus
Fenna Sille
12.
13. 38 YEARS
LONG LATENCY: Lived in Antofagasta during
the high exposure period
Lung cancer
OR = 4.35 (2.57-7.36)
No confounding by age, gender, mining work, SES, smoking, diet
Steinmaus et al. Cancer Epidemiol Biomarkers Prev. 2013 22(4):623;
Ferreccio et al. Am J Epidemiol. 2013 1;178(5):813-8
Kidney cancer
OR = 11.1 (3.60--
34.2)#
Transitional cell
Bladder cancer
OR = 6.88 (3.84-12.3)
16. Highly exposed in
utero and early life
Bornhere
40+ YEARS
Long term impacts of early-life exposure: Antofagasta
Bornhere
Cancer?
Highly exposed
as adults
17. BLADDER CANCER ODDS RATIOS (OR)
Steinmaus., Cancer Epidemiol Biomarkers Prev. 2014 23:1529-38
Cancer odds ratios comparing subjects exposed > 800 ug/L to those
exposed ≤110 ug/l by each age of exposure
19. Smith et al. Environ Health Perspect. 2012;120:1527-31
Ecologic Mortality Study of Early life Exposure: SMRs for
Antofagasta for subjects born during or just before the high exposure period
22. Born
BMI
age 20
BMI
age 40
BMI 10 years
preceding
interview
¿Cuál es su talla de adulto? _____”_____”
Mtro. Cms.
¿Cuál ha sido su peso más frecuente en su vida de adulto?
(En el caso de las mujeres no incluir peso durante el embarazo)
a. en los últimos 10 años________Kgr.
b. a los 20 años _________Kgr.
c. a los 40 años (si es > de 590 años) _____Kgr.
AGE
23. Odds ratios for arsenic and lung and bladder cancer:
People with low BMI (<90th
percentile)
24. Odds ratios for arsenic and lung and bladder cancer:
People with high BMI only after age 40
25. Odds ratios for arsenic and lung and bladder cancer:
People with high BMI only at age 20, not after
26. Odds ratios for arsenic and lung and bladder cancer:
People with high BMI at age 20 and after
32. Biologic plausibility
Arsenic is an effective treatment for promyelocytic leukemia
In vitro experiments with cancerous and noncancerous breast cells
36. Transgenerational impacts in humans: in Antofagasta?
1958-70
High exposure
F0
F2
F1
F3
2016
Ages 0-22 years
cancer
cancer
???
???
37. Summary
Ecologic and case-control studies
Who is most susceptible? Obesity and early-life exposure
Breast cancer: complimentary epidemiologic and laboratory studies
Public Health Relevance
Current regulations do not incorporate susceptibility factors
Mechanisms: surveillance, prevention, screening, treatment
Where To From Here?
Epigenetics: current studies with Northwestern University (DNA
methylation) and NCI (mRNA)
Multi- and transgenerational?
Other susceptibility: stress?
Exhaled breath study in Antofagasta
Superfund: hypertension, diabetes, prostate cancer, stress in Region II
Other chemical exposures in northern Chile?
38. Research Team
Catterina Ferreccio
Allan H Smith
Martyn Smith
Guillermo Marshall
Johanna Acevedo
Yan Yuan
Sandra Cortes
Luoping Zhang
Fenna Sille
Jane Liaw
Viviana Durán
John Balmes
Susana Cuevas
Lee Moore
José Garcia
Ken Cantor
Rodrigo Meza
David Kalman
Rodrigo Valdés
Roxana Parra
Gustavo Valdés
Vania Villagra
Hugo Benitez
Francisca González
Teresa Barlaro
Vania VanderLinde
Juan José Aguirre
Maria Isabel Vásquez
Liliana Pérez
Jacqueline Calle
NIEHS P42ES04705,
R01CA129558, R01ES014032
Notes de l'éditeur
I’ll use examples from our studies in Chile and elsewhere to discuss: 1. commonly used study designs in environmental epidemiology; the strengths and weaknesses of these designs, and what I see as some of the future directions in this field
and where our research is heading
Provide a little introductory information. Arsenic is a toxic element and there are a number of different exposure sources including…
Arsenic in water causes a variety of health effects including cancer, heart disease, and skin lesions, and much of what we know about the toxicity of arsenic comes from research done in northern Chile, specifically Region II. This area has a number of very important features for doing arsenic research. First, because there is so little rainfall, there are few individual drinking water sources, with most water coming from a small number of rivers that flow down from the Andes Mountains. This is in contrast to other arsenic exposed areas worldwide (US, Bangladesh, Taiwan, India) where most water comes from small private wells, that is, wells used by only a few people or families. In these areas there are thousands of wells and arsenic levels can vary dramatically from well to well. Because of this, in order to have a good idea of the exposures in these populations arsenic must be measured in thousands or tens of thousands of different wells, a task that is incredibly time consuming and expensive, and usually impossible.
Another major advantage for doing arsenic research in northern Chile is that arsenic levels have been measured in all the major water sources, with many records available from the distant past. This chart shows the arsenic levels in the larger cities and towns in Regions I and II since the 1950’s. Because we have such good records here records we simply need to know what city a person has lived in during what years to have a good idea of that person’s arsenic exposure, even from the distant past. This is important because as I will show you, we’ve found that these past exposures are the most important determinants of arsenic-associated cancer risks.
An advantage of studying arsenic in this area is the distinct high exposure period that occurred in Antofagasta. In 1958 two rivers with high arsenic concentrations were diverted to the city for drinking. Then in 1970 an arsenic treatment plant was installed. Overall, this lead to a 13 period of high exposure in the city. Because there was no other major water source in the city, essentially everyone was exposed. This unique exposure scenario has allowed us a rare opportunity to examine the long-term impacts of being exposed in utero or in early childhood. That is, by looking at the health risks now, in people who were born during the high exposure period, we can assess the specific impacts that early-life exposure may have on the risks of adult diseases like cancer, diabetes, and hypertension.
Our first studies in this area were “ecological”, that is, studies in which we simply compared the overall cancer rates in Region II to those in the rest of Chile. In this study design we only had data on whether or not people died of cancer and where they lived at the time of death. When looking at these data we found very high rates of bladder cancer, lung cancer, and skin cancer mortality in Region II compared to the rest of the country. This work was done in collaboration with Drs. Marshall and Ferreccio at PUC.
The previous slides I showed on arsenic were from ecologic studies. Again, we only had information on the rates of death and the cities where the people lived at the time of death. We did not have data on their long-term residential history, where they were born, or data on potential confounders like smoking, diet, or other exposures.
To respond to these criticisms we applied to and received funding from the National Institutes of Health in the US to do research in which we did collect individual data on arsenic exposure and potential confounding variables like smoking. Here we chose the case-control study design. This began with establishing a rapid case ascertainment system in which we identified all incident cases of bladder, lung, and kidney cancer in Regions I and II for the years 2007-2010. For a comparison group, we randomly selected people without these cancers from the Chile Electoral Registry, matched to cancer cases by age and gender. Each subject was then interviewed regarding their full residential history, smoking history, jobs, and other potential confounders.
Each residence the subjects lived in throughout their lives was then linked to arsenic water records we had available for all the cities in Regions I and II so that we could assign an arsenic level to each year of each subject’s life, beginning at birth and ending at the time of interview.
We assembled a team of Chilean and US researchers including…
Under the leadership of Dr. Ferreccio we established a research infrastructure that included offices, laboratory facilities, research staff, and collaborators in the four largest cities in Regions I and II.
Here are our findings. First, we focused on people who had lived in Antofagasta. This was done because this city had a very high arsenic exposure period beginning in 1958 when two rivers with high arsenic concentrations were diverted to the city for drinking, and ending in 1970 when an arsenic treatment plant was installed. As in our ecologic studies, we found very high rates of lung, bladder, and kidney cancer almost 40 years after the high exposure period ended.
The prolonged high risks we saw for arsenic is unusual. For other carcinogens, like tobacco smoke, the risks of cancer drop fairly rapidly after the exposure ends. With arsenic we are finding something different: cancer risks are remaining high many years after the high exposures have ended. This highlights the importance of having data on exposure from the distant past when examining some environmental toxins like arsenic. That is, if we want to assess the true risks of certain environmental contaminants, we may need to examine exposure patterns over a period of decades.
More and more research is not just focusing on whether or not environmental chemicals cause disease, but also if they do cause disease, who is most susceptible. This is important, because most current environmental regulations are based on the toxic effects seen in “normal” healthy individuals. If some groups are more susceptible to these chemicals then others, then these current regulations may not be adequately protective in these particular groups.
An advantage of studying arsenic in this area is the distinct high exposure period that occurred in Antofagasta. In 1958 two rivers with high arsenic concentrations were diverted to the city for drinking. Then in 1970 an arsenic treatment plant was installed. Overall, this lead to a 13 period of high exposure in the city. Because there was no other major water source in the city, essentially everyone was exposed. This unique exposure scenario has allowed us a rare opportunity to examine the long-term impacts of being exposed in utero or in early childhood. That is, by looking at the health risks now, in people who were born during the high exposure period, we can assess the specific impacts that early-life exposure may have on the risks of adult diseases like cancer, diabetes, and hypertension.
One potential susceptibility factor is age of exposure. In our Chile study, because we had data on lifetime exposure we could examine this issue. Here we looked at risks by age of exposure. These figures show the lung and bladder cancer odds ratios comparing high to low arsenic exposure groups for each age at which the exposure occurred. As can be seen, for both lung and bladder cancer, the odds ratios of cancer were higher when exposures occurred at younger vs. older ages. Data like this highlight the importance of having exposure data not just at one point in time but also over the entire period of a person’s life.
Lung cancer had a slightly different pattern. Risks were still higher with earlier vs. later exposures, but the high risks continued for exposures throughout childhood. This period corresponds to the period of lung development (ages 0-20 years) suggesting that any time while the lung is growing is a period of potential susceptibility
First set of studies
Ecologic analyses of mortality
Comparing Region II (Antofagasta) to the rest of Chile
No info on migration (bias to the null)
No info on confounders like smoking
Next I’d like to talk about obesity. The topic of obesogens has received considerable attention recently. These are chemicals that are thought to increase obesity, and a variety of chemicals like pesticides, BPA and phthalates in plastics have been implicated (although data are somewhat inconsistent).
In our Chile study, we’ve looked at obesity in a different way. Instead of looking at it as an effect of arsenic, we examined it as a potential susceptibility factor. Our decision to look at it this way was based on findings that both arsenic and obesity increase cancer, and both increase factors such as inflammation and oxidative stress that are also linked to cancer.
Our questionnaire included questions about body mass index (BMI) for ages 20, 40, and for the 10 years before interview
We had data on BMI at three points in time. Here is what we found. In people who had low BMI at all three time points we assessed, the risk (or odds) of cancer increased with increasing arsenic exposure, but only slightly (up to an odds ratio of about 3 at the highest arsenic exposure)
In people who only had an elevated BMI at age 40 or afterwards (but not at age 20), the risks for arsenic-related cancer increased a little bit more
In people with elevated BMIs at age 20, but not later, risks were even greater
The strongest arsenic-cancer association was see in people with elevated BMIs both early (age 20) and later, with cancer odds ratios approaching 30.
Another expanding area is the combination of basic epidemiology studies with laboratory studies aimed at helping to identify the mechanisms for the epidemiologic findings that are seen. An example of this is our work on arsenic and breast cancer.
In this study we compared breast cancer rates in high arsenic Region II with those in Region V, a sociodemographically similar region with very little arsenic in drinking water. Before the high exposure period in Antofagasta (before 1958 when the high exposures started), the Region II and V breast cancer mortality rates were similar (relative risk = 1)
However, during the high exposure period (1958-70), breast cancer rates dropped dramatically: about 50% overall, and about 60% drop for breast cancers in women under age 60.
Since the high exposure period ended, the rates of breast cancer in Region II have begun to rise, slowly back up to those in Region V.
The interesting features of these findings are that the rates dropped so much, and that they almost perfectly tract the high and low arsenic concentrations in Antofagasta. This makes it highly unlikely that some factor other than arsenic was the cause of this drop. In interviews we have had with physicians and public health officials we can find no reason to believe this drop is related to differences in screening, treatment, or known breast cancer risk factors.
To help verify our epidemiologic findings we worked with laboratory scientists at Stanford University who dosed normal breast cells (black diamond) and breast cancer cells (others) with varying levels of arsenic. They found that some breast cancer cell lines (especially SKBR3 and MDA468) were killed at lower arsenic doses than normal (non-cancer) breast cells, suggesting that arsenic may preferentially kill breast cancer cells. This is an example of how bio-molecular laboratory data can be used to help add plausibility or clarity to new epidemiologic findings.
Another topic that is receiving increasing attention in environmental epidemiology is the transgenerational effects of chemical exposures. In other words, can the adverse effects of environmental exposures be transferred across multiple generations so that your health is not just determined by your exposures but also those of your parents and grandparents. F0 = the exposed pregnant animal, F1 is the fetus, and F2-3 are subsequent generations.
If this does occur, the most likely mechanism would be through epigenetics, and the most likely critical exposure period would be the fetal or early-life periods since these are times of major epigenetic re-structuring of the human genome.
Epigenetics: factors like DNA methylation that controls the expression of genes
DNA methylation may be a target site of early-life and fetal exposures because these early periods are periods of major reorganization of DNA methylation patterns in the human genome. As shown here, there is a major demethylation and re-methylation of the human genome in the developing gametes and in the early embryo.
In fact, these types of transgenerational effects have been seen in animal studies. In this study, vinclozolin (a toxic pesticide) was given to pregnant rats whose offspring were then followed for up to four generations. As seen here, evidence of adverse impacts on sperm were not only seen in the immediate offspring (F1 generation) of the exposed rats, they were also seen up to four generations later, that is in the exposed rats grandchildren (F2), great grandchildren (F3), and great-great grandchildren (F4).
Leads to the question: could we do a study of arsenic and transgenerational effects in northern Chile? In other words, our studies todate have involved people who were directly exposed during the high exposure period, either as an adult (F0) or as a fetus (F1). To date we have found high cancer rates in F0 and F1. But what about these peoples children or grandchildren? The F3 generation would have never been high highly exposed. But if adverse effects can be transferred through epigenetic effects in their grandparents, these ffects could potentially influence their health. There are few situations in the world like Antofagasta: where we have a very well documented period of high exposure, in the distant past, in a very large population, to a well known highly toxic agent like arsenic. This unusual scenario would be a rare and perfect situation to study to confirm the animal findings and see whether transgenerational toxicity may actually occur in humans.