Chronic Arsenic Poisoning:
History, Study and Remediation

Site History
Harvard/MIT Faculty
Searchable Address List
Online Articles
Pictures of Sufferers
Last modified May 18th 2011
Link to a translation into Rumanain at

<a href="">
I. Chronic Adverse Effects

    After a few years of continued low level of arsenic exposure, many skin ailments appear, i.e. Hypopigmentation (white spots), Hyperpigmentation (dark spots), collectively called Melanosis by some physicians and dyspigmentaion by others.  Also keratosis
(break up of the skin on hands and feet).   We have assembled a collection of photographs of patients with typical symptoms.   We have plots of incidence of these ailments versus integrated dose and versus maximum dose for three villages in Inner MongoliaThese suggest lineraity of incidence with dose above a possible threshold at 75 ppb in the water,  with incidence rising to approach 100%  near  1000 ppb.  In developed countries dyspigmentation and even keratosises do not always lead to death.  But in Bangladesh, a farmer with keratoses on his feet may continue to walk upon them and develop gangrene.  Then his foot must be amputated and he can no longer work and feed his family.    So far several thousand cases of melanosis have been identified in Bangladesh but it is important to note that as of July 2005 only about 30% of the villages have been surveyed. More information from greater number of surveys may show a much greater number of  people affected.

    After a latency of about 10 years, skin cancers appear. After a latency of 20 - 30 years, internal cancers - particularly bladder and lung appear.These have all been seen in Taiwan and in Chile.  The data from Taiwan and Chile suggest that biggest risk of death is from internal cancers. The data from Chile suggests that 30 years after a 5-year exposure of drinking water at a concentration of 0.5 ppm (and reduced exposure thereafter) will result in a cancer risk of 10%. No one knows how to extrapolate to the lower exposure of 0.05 ppm (50 ppb or 50 micrograms per liter), but if one takes a "default" linear response,   the risk is 1%.   At the WHO guideline of 10 ppb, to be effective in the USA about 2006, the risk is still 0.2%.

     Arsenic absorption is also believed to lead to vascular disease.    A medical glossary also lists symptoms.   Other comments upon chronic health effects come from the University of Minnestota

Estimate of Deaths in Bangladesh

<>    The number of cancers expected in Bangladesh from the exposure already undergone can be very roughly estimatedby using this "default" assuming that there are 20,000,000 to 70,000,000 exposed persons at levels between 0.05 ppm and 0.5 ppm. These estimates will vary between 200,000 and 2,000,000. Correcting for other sources of death would reduce this. The life expectancy in Bangladesh is lower than in Chile and in Taiwan and this would also lower the result.   Even the lower estimate of cancers anticipated in Bangladesh exceeds ten fold the estimate of cancers (20,000) that might, on the  pessimistic linear dose response calculation, be caused worldwide by the 1986 Chernobyl catastrophe.   This has led the webmaster to say on the BBC in UK and on National Public Radio (NPR) in USA that the catastrophe in Bangladesh makes Chernobyl look like a Sunday School picnic.

Scientific Questions

    There are a number of interesting scientific questions that can be raised and may be answered at some time in the future.    For other chemicals, laboratory animals have often given indications on what we should expect, but since it has proven hard to persuade laboratory animals to get cancer or other lesions at equivalent doses, the applicability of animal tests is highly questionable.
    (1)   Are dyspigmentation, keratoses, skin cancers, and internal cancers all different distinct medical end points or is there a natural progression from dyspigmentation, through keratoses to skin cancer?   There are indications that Dyspigmentation in its early stages can be reversed on provision of pure water, but that keratoses may be irreversible.   Also that skin cancers appear at locations on the skin where there has been some skin damage.
    (2)   Is there a threshold below which arsenic damage to people does not occur?  Is such a  threshold the same for all lesions, or is it just for skin lesions, or for the uncommon cancers such as bladder cancer?  This is a difficult question to answer. This was discussed in the  framework of arsenic induced lesions by Wilson in a poster at the 4th International Arsenic Conference in San Diego (and introduced to EPA)    There are a number of subsidiary questions which this discussion raises.
    (2a)    Is the induction of any or all arsenic lesions a deterministic or a stochastic response?   Not all members of a population exposed to a moderate level of arsenic develop lesions.  Attempts to understand any differences between individuals that could explain differences in response seem to have had no success.  Therefore it is usually assumed that the induction of these lesions is a stochastic rather than deterministic.   Yet some evidence from Inner Mongolia presented by Tucker et al. in posters at the 5th International Arsenic Conference suggest that there may be more determinism than often assumed.  Using the criteria for dyspigmentation employed by the Huhhot Anti-Epidemic Station the data suggest a threshold in response at about 75 ppb.  Yet when Tucker's more sensitive method is used dyspigmentation is seen below 75 ppb suggesting that any level of arsenic produces an effect that develops deterministically into more serious effects.
    (2b)    Does arsenic act indirectly, and differently from background,  on a stage of the process with a non linear process?  The general default argument for low dose linearity depends upon the usual  assumption, that cancer is caused by a multistage process,  and arsenic is assumed to act directly on one stage in that process in the same way that natural, background, processes act.   Then, as shown by many authors over the years (Doll, Peto, Crump. Crawford, Wilson, Garwin) low dose linearity becomes a simple mathematical consequence  (Taylor's theorem).   Is that a correct assumption?  Moreover the linearity suggested by Taylor's theorem is only at low doses.  Common sense suggest that low means appreciably less than the natural processes.  It might well apply at the doses of concern for lung cancers which are common, especially among smokers,  but does it apply for bladder cancer whose incidence at 500 ppb seems 10-50 times more than the natural incidence?
    (2c)    Arsenic is not a genotoxic agent.  Does this fact make any difference to the discussion of linearity in question (2b)?  Clearly only a genotoxic agent is expecte to act at the first, initiating, stage of the cancer process.   But all agents, can act at later stages.  It is noteworthy that several non-genotoxic agents, asbestos, benzene and dioxin are not genotoxic yet are assumed for regulatory purposes to have a linear dose response.  Should arsenic be treated any differently from these, and if so, why?
    (2d)   Can studies of large populations exposed at low levels tell us much?    Many people say no. There are too many problems that lead to variations between groups larger than statistical fluctuations suggesting that even taking a larger study will not give a significant effect.   But in spite of the well known probelsm with ecological studies, Lamm et al. presented at the 5th International Arsenic Conference a study of the variation of bladder cancer across seveal US counties with different arsenic levels.  Up to 50 ppb in the water no relation with arsenic level was seen, and the largest effects considered possible by the US National Academy studies could be excluded.  The variation between counties was large (with a fluctuation of 23%)  which suggests that if we could understand the reasons (or some of them) for the variation, a much stronger statement could be made.
    (3)  Are there any remedial actions that can cure arsenic poisoning or alleviate its symptoms?  Answering this question is a major aim of several epidemiological studies including the one being carried out by Professor Christiani at the Harvard School of Public Health.    From general considerations one might expect, for example, that arsenic related cancer incidence could be reduced a factor of two by a good diet (fresh fruit and vegetables) as compared to a bad diet.  Even cigarette related lung cancers are so reduced in US studies.   As of summer 2005 it has been shown (PhD thesis from HSPH) that ingestion of betel nuts increases the sensitivity by a factor of 2.    It has been recently claimed that the medical algae Spirulina  helps the treatment of  of patients with chronic arsenic poisoning.  Selenium is a chemical antagonist of arsenic and many people have suggested that it be used.  But over doses of selenium can cause troubles themselves and there is general agreement that the use of selenium requires close supervision.


II   Beneficial Uses of Arsenic
    Although arsenic has been known as a poison for thousands of years there have long been reports of beneficial effects.  
Arsenic has been used for over 200  years for  medicinal purposes.   In a 1% solution, it was recommended in 1792 by Dr Folwer of Edinburgh and Fowler's solution was in the British pharmacopeia for many years as a remedy for stomach ailments and several digestive problems
<><>.   For occasional use the dose integrated over time was small and no adverse outcome was observed.  But some sufferers took Fowler's solution for a year or moe.  Hutchinson in 1888 described skin lesions that occurred as a result.  Arsenic used to be used as a cure for diseases  such as syphilis before other antibiotics were found to be more effective.     When used externally, arsenic is still used for its supposed beneficial properties.  Many mineral hot springs have high levels of arsenic.  For example, at Ojo Caliente in New mexico, one can choose between bathing in an arsenic pool, an iron pool or a sodium pool.  While testing this out perosnally, this webmaster was told bby several bathers in the arsenic properties of their conviction that arsenic was doing them good.   Others have written about other possible beneficial effects at very low levels. It is important to note that the beneficial effects are for different medical outcomes (end points) than either the acute or chronic adverse effects and that both beneficial and adverse effects  an be observed simultaneously (as is well known for alcohol ingestion).    Another detailed article about beneficial uses of arsenic can  be found here .

    The recent article  " Complete Remission After Treatment of Acute Promyelocytic Leukemia with Arsenic Trioxide" shows that moderate to high doses of arsenic given fora period of 30 days can cure leukemia.  It is important to realizethat this is not in contradiction to the evidence that arsenic given atlow doses for 15 years or more can produce high cancer rates (i.e.. kidney,bladder, lung).  
The doses given for leukemia cures were for about 30 days and the accumulated doses were about 300 mg.  By way of comparison a person drinking  2 liters of water a day for20 years with 500 µg/litre of arsenic (as was the case in Chile) will accumulate 7500 mg - 20 times more.  So if the accumalated dose is the criterion for developing cancer, as is strongly believed, there should be no problem with the doses given for curative purposes.

    An abstract of this article and abstract (in English)of another article (in French) commenting upon it are available here.

III  Animal Studies
    That arsenic at low levels is safe seemed to be reinforced by animal studies that seemed to show that arsenic is beneficial (to animals) at low doses. Indeed, the fact that laboratory animals could not be persuaded to develop cancer misled toxicogists  throughout the world and greatly contributed to the present catastrophe.
 There are a number of studies of  effects of arsenic on laboratory animals.  However, most of them do not show effects at the levels that are showing effects in people.  This remains a huge puzzle. The webmasters would appreciate input from those that understand this issue.

The work of Dr Toby Rossman seems to be one of those that may help illuminate the problem.   In particular she has shown that arsenic works as a co-carcinogen with ultraviolet light in causing skin cancers.   This has also been shown by Tucker and collaborators.

Contact  us

Please tell us what is missing from this website, and send any useful material you know about to us at .   If you wish to be informed of any updates, please send an e-mail and say: "Please keep me informed!"

Webmasters: Richard Wilson

since June 26th, 2003