Effects and dose--response relationships of skin cancer and blackfoot disease with arsenic.

In a limited area on the southwest coast of Taiwan, where artesian well water with a high concentration of arsenic has been used for more than 60 years, a high prevalence of chronic arsenicism has been observed in recent years. The total population of this "endemic" area is approximately 100,000. A general survey of 40,421 inhabitants and follow-up of 1,108 patients with blackfoot disease were made. Blackfoot disease, so-termed locally, is a peripheral vascular disorder resulting in gangrene of the extremities, especially the feet. The overall prevalence rates for skin cancer was 10.6 per 1000, and for blackfoot disease 8.9 per 1000. Generally speaking, the prevalence increased steadily with age in both diseases. The prevalence rates for skin cancer and blackfoot disease increased with the arsenic content of well water, i.e., the higher the arsenic content, the more patients with skin cancer and blackfoot disease. A dose-response relationship between blackfoot disease and the duration of water intake was also noted. Furthermore, the degree of permanent impairment of function in the patient was directly related to duration of intake of arsenical water and to duration of such intake at the time of onset. The most common cause of death in the patients with skin cancer and blackfoot disease was carcinoma of various sites. The 5-year survival rate after the onset of blackfoot disease was 76.3%; the 10-year survival rate was 63.3% and 15-year survival rate, 52.2%. The 50% survival point was 16 years after onset of the disease.

In a limited area on the southwest coast of Taiwan where artesian well water with a high concentration of arsenic has been used for more than 60 years, a high prevalence of chronic arsenicism has been observed in recent years (1)(2)(3). It is well known that hyperpigmentation, keratosis, and cancer are the major manifestations of chronic arsenicism from any source (4), but peripheral circulatory disorders have also been reported occasionally in chronic arsenicism (5)(6)(7)(8)(9). It seems reasonable to assume that arsenic may be the common etiological factor for skin cancer and blackfoot disease. A positive association between the arsenic level of drinking water and the prevalence of skin cancer in endemic areas of chronic arsenicism has been reported from the district of Reichenstein, Silesia, Poland (7), Cordoba Province, Argentina (10), and most recently from Antofagasta, Chile (11). Although reliable information about the number of cases of arse-*Department of Medicine, National Taiwan University, College of Medicine, Taipei, Taiwan, Republic of China. nical cancer of the skin is Reichenstein is not available, Arguello and Tello identified a total of 148 cases in Cordoba Province (12). There were no statistical studies in either population group.
After a preliminary survey of the endemic area in Taiwan, we have learned that a high percentage of the inhabitants suffered from chronic arsenicism, and a considerable number had arsenical skin cancer or blackfoot disease, a folk term for a peripheral vascular disorder resulting in gangrene of extremities, especially the feet (1)(2)(3). The affected area afforded us a unique opportunity to study systematically the epidemiologic, clinical, and pathologic aspects of both these diseases (Figs. 1 and 2).

Population of the Endemic Area
We use the term "endemic" area to refer to the total region where artesian wells providing water with a high concentration of arsenic have been used for a long time or are still used, and where a high percentage of inhabitants suffer from skin lesions associated with chronic arsenicism, i.e., hyperpigmentation, keratosis, and cancer. The total endemic area thus defined consists of two districts (Peimen and Yi-Chu), three towns (Putai, Hsueh-Chia, and Yen-Hsui), and three villages (two from the suburb of Tainan City and one from the An-Ting district). Census records obtained from villages in the endemic area showed that the entire population at risk is 103,154 (51,289 males and 51,865 females), although approximately 10% of the people are constantly away from home, either for employment or for other reasons. The inhabitants are mostly engaged in farming, fishing, or salt production. The socioeconomic state of the people is poor. As is usually true in rural areas in southern Taiwan, the people subsist on food low in protein and fat; carbohydrate, rice, and sweet potatoes constitute the main part of their diet. The customs and habits of the inhabitants do not differ from those in other parts of Taiwan.

Artesian Wells
Artesian wells have been in use since the decade 1900-1910. In most parts of the area where artesian wells are still in use, they are 100 to 280 m deep, 80% being between 120 and 180 m in depth. In 1956 water was piped to many places mostly from the reservoir of the Chia-Nan irrigation system, this water having an arsenic content of 0.01 ppm. In 1966 a tap water supply was made available to the majority of the endemic area in Tainan County. There will be no more use of artesian wells in the study area. In the villages surveyed, the arsenic content of the well water ranged from 0.01 to 1.82 ppm, including four shallow wells (20). The water from the artesian wells was not always as high in arsenic content, for Environmental Health Perspectives some from the neighborhood of the endemic area had an arsenic content of only 0.01 ppm. Shallow wells were usually almost free from arsenic (0.001 ppm).

Methods of Survey
By the end of 1965 a general survey of the entire population of 40,421 in the endemic area had been completed. Ten of the 37 villages are in Chai-yi County, 25 in Tainan County, and 2 in a suburb of Tainan City.
The survey was house-to-house and every member of the family was examined. Particular attention was paid to skin lesions, i.e., the pigmentation of unexposed body surfaces, keratoses (especially on palms and soles), and cancers. In addition, the blood pressure of persons over age 20 was taken; pulsation of the dorsalis pedis and posterior tibial arteries was noted; evidence of any peripheral vascular disorder was recorded. The clinical examinations were done by two young doctors assisted by a nurse and clerk, and all lesions they found were checked by one or two of the experienced senior members of our staff. A total of 303 arsenical skin cancers in 184 patients were studied histologically. These lesions are discussed in detail elsewhere (21). From the villages surveyed, we examined 142 water samples from 114 wells, of which all but four were artesian (Fig. 4). The villages in August 1977 ill the endemic area that still used artesian wells at the time of examination and had water with an arsenic content over 0.01 ppm are shown in Figure 3. The method used for determination of the arsenic content of the water was that of Natelson (22). 20

Method of Follow-Up Study of Blackfoot Disease
A total of 1108 patients with blackfoot disease in the entire endemic area were examined by the author himself during 1958 and 1975. There 669 males aaid 439 females. Each patient was carefully studied. Criteria for diagnosis depend upon (1) objective signs of ischemia, i.e., absence or diminution of arterial pulsations, pallor on elevation or rubor on dependency of ischemic extremities, and various degrees of ischemic changes in the skin and (2) subjective symptoms of ischemia, i.e., intermittent claudication, pain at rest, and ischemic neuropathy. Follow-up was attempted by a variety of methods, including direct follow-up physical examination, several mailings to patients, and search of death certificate files. By the end of the follow-up period 528 patients had died, giving an overall fatality rate of 47.7%. A history of typical ischemic symptoms such as numbness or intermittent claudication antecedent to Lhe ulcer or gangrene has been used in estimating the date of the onset of blackfoot disease. The duration of intake of arsenical water at the time of onset represents the period of time between first use of such intake and the time of onset of the disease. Duration of intake of arsenical water represents the duration of time during which the patient started drinking artesian well water up to the time of survey or up to the time of change of source of drinking water. For native patients the duration is estimated to be equivalent to their ages, but for the patients who came from areas without artesian wells the duration was counted as starting from the year of their arrival.
Evaluation or rating of permanent disability has long been recognized as an important and complex subject. The percentage of permanent impairment due to amputation has been taken from the literature (23).
The original classification did not have 1-4% of the impairment of the whole man, but many minor amputations occur in mild cases of Blackfoot disease. For this reason an additional class intermediate between 1 and 2 was added to represent very mild impairment, 1-4% of impairment.

Age-Specific and Sex-Specific Prevalence of Skin Cancer
Of the population of 40,421,428 (10.6/1000) had arsenical skin cancer (Table 1). There were no patients under 20 years of age. The prevalence rate increased markedly with age, except for females above age 70. Over 10% of the people above age 60 were affected by skin cancer. The overall male-tofemale ratio was 2.9 to 1, with males having a higher rate in all age groups above 30 years. The villages surveyed were arbitrarily divided into three groups, according to the arsenic concentration in the well water being designated "low" (below 0.3 ppm), "mid" (0.3-0.6 ppm), and "high" (above 0.6 ppm) groups. A clear-cut ascending gradient of prevalence of skin cancer from low (L) to mid (M) to high (H) groups was found for both sexes in the three age groups: e.g., 20-39 yr (H 11.5, M 2.  Figure 5. Another group, "undetermined," included those villages where either artesian wells with arsenic-polluted water were no longer in use, or the difference in the arsenic content in water from various artesian wells in the same village was so great that it was impossible to put them into any of the above-mentioned classifications. The prevalence rate of the "undetermined" group was similar to that of the "mid" group.

Age-Specific and Sex-Specific Prevalence of Blackfoot Disease
The overall prevalence was 0.9-1.1% for males and 0.7% for females (Table 2). After age 40, the rates for males were significantly higher than for females. The prevalence rose steadily until age 70 for both sexes; after that, rates decreased. The overall male-to-female ratio was 1.3 to 1. According to the same above-mentioned classification (low, mid and high groups) in the villages surveyed, the prevalence rate for blackfoot disease also revealed a clear-cut ascending gradient from low to mid to high groups for both sexes in the three age groups: e.g., 20-39 yr (H 14.  Figure 6. The greater the arsenic content, the higher the prevalence of blackfoot disease.

Combination of Skin Cancer and Blackfoot Disease
In our survey of chronic arsenicism in a population of 40,421 in 37 villages, we found 428 cases of skin cancer and 370 cases of blackfoot disease giving an observed rate per 1000 of 10.59 and 8.96, respectively. The combination of skin cancer and blackfoot disease occurs in 61 cases, or 1.51 per 1000, whereas 4 cases, or 0.09 per 1000, are expected. Thus the ratio of observed to the expected is 16.77. This ratio indicates that the coincidence of the two conditions cannot be attributed to chance, and incriminates a common or underlying factor, namely, chronic arsenicism.

Relation of Water Supply to Blackfoot Disease
It was determined that all of the patients with blackfoot disease had consumed artesian well water before the onset of the disease and none of the residents of the endemic area who had consumed only surface water or water from shallow wells developed blackfoot disease. This appears to be because the shallow well water is almost free from arsenic (0.001-0.017 ppm). In some parts of these areas, a tap water supply was provided in 1956. The change in the frequency of blackfoot disease since that time is strong evidence supporting the arsenic hypothesis. As shown in Table 3, in all periods studied-1955 and before, between 1956 and 1965, and between 1966 and 1975-there was no statistical difference in the age distribution between the districts with water supplied by artesian well and artesian well with a changeover to tap water in 1956. There were, however, more young patients in the A group than in the B group during the period 1956-75. It should be particularly noted that while there were no cases less than 20 yr old present in the B group, there were five cases in this age group in the A group during 1956-75. This fact is most important, since it shows that no cases were found among the inhabitants who were born after the tap water supply was established. These supplemental data support the conclusion that a close association exists between the consumption of arsenical water and the development of blackfoot disease.

Hypothesis of the Cause of Blackfoot Disease
The hypothesis is: the occurrence of endemic blackfoot disease is directly related to the arsenic content of artesian well water.
The patients with blackfoot disease were classified into one of the three groups according to the arsenic concentration in the well water: high, mid, and low groups. A definite gradient was associated with the degree of exposure to arsenic (Table 4). Furthermore, with each cohort (duration of intake of arsenical water) a gradient in blackfoot patients was associated with degree of arsenic concentration. During of intake of arsenical water and degree of arsenic concentration were directly related and the difference was statistically significant (X2 = 11.53, n = 4, 0.025 < p < 0.01). Thus exposure to arsenic was found to be associated with an excess of frequency of blackfoot disease in the high exposure group of longer consumption of artesian well water. Thus, the hypothesis is sustained.  (Ix2 = 11.53, n = 4, 0.025 < p < 0.01. 25-49% groups decreased in the 20-39 yr period of intake of arsenical water at the time of onset. From this, it can be supposed that this is a chronic type in which disease goes on gradually. Thereafter cases with both 25-49% and 50-74% impairment increased tremendously at -40 yr intake of arsenical water prior to time of onset. In general, the degree of permanent impairment of patient is significantly correlated with duration of intake of arsenical water at the time of onset (X2 = 55.03, n = 10, p <0.001).

Degree of Severity in Relation to Duration of Intake of Arsenical Water at the Time of Onset
As an aside, it may be noted that there is an acute type of reaction; the patient develops (severe impairment suddenly after a prolonged period of intake of arsenical water prior to the time of onset. Generally speaking, group in which the duration of intake of arsenical water at the time of onset was below 19 yr did not correlate with the permanent impairment in the 0%, 1-4%, 5-24%, and 25-49% impairment groups. Those account for 24.8%, 16.5%, 24.8%, and 26.4%, respectively. This fact indicates that there is an acute type which develops rather severe impairment in a shorter duration of intake of arsenical water at the time of onset. Degree of Severity of Blackfoot Disease Related to Duration of Intake of Arsenical Water As shown in Table 6, it is also true that the degree of permanent impairment is, in general, closely cor- related with duration of intake of arsenical water, especially in severe cases. The percentage of impairment in 60 yr is much lower than that of the patients ingesting arsenical water for less than 40 yr. On the contrary, the percentage of moderate impairment increased with longer duration of intake of arsenical water. The differences of percentage of permanent impairment among different period of intake of arsenical water were statistically significant (X2=34.17, n=15, p<0.005). Thus the degree of permanent impairment of patients with blackfoot disease was noted to be directly related to duration of intake of arsenical water.

Mean Age at Death of Patients with Blackfoot Disease and Arsenic Level in the Well Water
Of the 528 deaths, for 379 patients with blackfoot disease, the mean age of death was found to be 58 + 12.6 yr, 60.0 + 12.7 yr, and 63.7 + 12.2 yr, respectively, for high, mid, and low As levels ( Table 7). The difference of mean age at death between high and low groups was statistically significant (p<O.Ol). The mean age of death for patients with blackfoot disease was significantly lower in areas of high arsenic content than in areas of low arsenic content of the well water. Case Fatality of Patients with Blackfoot Disease The relationship between age at onset and subsequent fatality is shown in Table 8. The greatest concentration of age at onset was in the 50-59 age group. The case-fatality rate increased linearly with age at the time of onset, with greater risk for older patients, reaching 66.3% by 70 yr or over, in general, the older the age group the higher the rate.

Causes of Death in Patients with Skin Cancer and Patients with Blackfoot Disease
During the follow-up period of 18 yr for blackfoot disease, 528 of 1108 patients died. This is an overall fatality rate of 47.7%. Another 10-yr follow-up period for skin cancer showed 244 of 428 patients died, giving a 57.0% fatality rate.
An analysis was made of the causes of death up to 1975 in patients with these two diseases as compared with that in the general population of the endemic area(( Table 9). The most common cause of death in patients with skin cancer or blackfoot disease was carcinoma of various sites, 27.9% in skin cancer and 18.8% in blackfoot disease. Cardiovascular disease was also responsible for 15.7% of death in patients with blackfoot disease. In general, in the population of the endemic area in 1966, cancer accounted for 13.1% of the deaths and cardiovascular disease 9.1%, whereas for the whole population of Taiwan in 1966 cancer and cardiovascular disease accounted for 7.9 and 8.1% of cases, respectively.
Death Rate of Patients with Blackfoot Disease Table 10 shows the death rate among the patients with blackfoot disease. An annual death rate per 1000, as might be expected, increased somewhat with age, 528 deaths in 12,461 patient-years in this series give an annual death rate of 42.4 per 1000. Survival rate represents the period of time between the onset of disease and the end of the study. Figure 7 shows the survival rate each year after the onset of disease for amputated, nonamputated, and total patients. The 5-yr survival rates after onset of blackfoot disease for total, amputated, and nonamputated patients were 76.3%, 73.5%, and 80.9%, respectively. The 10-yr survival rates were 63.3%, 60.0%, and 68.8%, respectively; 15-yr survival rate, Environmental Health Perspectives 52.2%, 48.2%, and 60.0%, respectively. The 50% survival point for total, amputated, and nonamputated patients were 16, 14, and 21 yr, respectively. The annual death rate per 1000 by the end of 15 yr was 51.6 per 1000 in amputated patients, 38.1 per 1000 in nonamputated patients, and 46.0 per 1000 in total blackfoot patients. It must be pointed out that many factors, such as duration of intake of arsenical water, may influence the results.

Survival Rate of Patients with Blackfoot Disease Related to Duration of Intake of Arsenical Water
It is also worthwhile to give special consideration to the relationship between the survival rate and the duration of intake of arsenical water. The duration of intake of arsenical water was divided into four groups; 1-19 yr, 20-39 yr, 40-59 yr, and >60 yr. As shown in Figure 8, there was a clear-cut descending gradient of survival rate from the >60 yr group to 40-59 yr group to the 1-19 yr group. The 50% survival points for patients with 1-19, 20-39, 40-59, and >60 yr intake of arsenical water were 25, 31, 14, and 5 yr, respectively. Annual death rates per 1000 by the end of 15 yr were 5.9 per 1000 in the 1-19 yr intake group 14.7% per 1000 in the 20-39 yr group, 50.8 per 1000 in the 40-59 yr group and 112.2 per 1000 in the >60 yr group.

Discussion
Although arsenic is notorious primarily for its acute toxicity, chronic toxicity is also a problem. Despite the many papers dealing with the medical, dietary, and occupational etiology of arsenical cancer, no reliable data on the frequency of arsenical cancer in a total population at risk are available in the literature. Arsenic is a common mineral, and August 1977 in many parts of the world it appears in drinking water obtained from wells drilled into arsenic-rich ground strata. The classical examples of toxicity from consumption of arsenic-polluted public drinking water are presented by the occurrence of skin cancer in members of some population groups, residing in the district of Reichenstein, Silesia, Poland (7), in Cordoba Province, Argentina (10) and recently in Antofagasta, Chile (11). Although a positive association between the arsenic level of drinking water and the prevalence of skin cancer in endemic areas of chronic arsenicism has been reported from the above-mentioned area, a negative finding was also noted in Lane County, Oregon, U. S. (24). The Lane County water arsenic levels averaged much lower than those reported from Taiwan and Antofagasta (11), so that they suggested that their increasing incidence of basal cell carcinoma might be potentiated by urban air pollutants.
Skin cancer, which is very common in Caucasians, has a relatively low incidence among Chinese in Taiwan-only 2.9% (25). Skin cancers usually occur on exposed surfaces, i.e., on the head, face, and extremities, and often are epidermoid or basal cell carcinomas. These features are also present in patients with ordinary skin cancers in Taiwan, but do not appear in the present series of arsenical cancer cases. The most common type of lesion (26) was intraepidermal carcinoma (51.7%), and the body areas most frequently involved were unexposed surfaces (74.5%). In addition, the extremely high percentage of cases with multiple skin cancer (99.5%) was characteristic of our series.
In general, the prevalence of skin cancer, hyperpigmentation, and keratosis increased steadily with age. As the inhabitants of the endemic area started using artesian wells for drinking water more than 55 yr ago, the period of exposure to arsenic-polluted water was very long in those who were over age 50 at the time of examination. It was difficult to elicit from patients the age at onset of arsenical cancer, because most of the patients were unable to name a date. We know from this study that the youngest cancer patient was 25, the youngest with hyperpigmentation was 5, and the youngest with keratosis was 15. This means that hyperpigmentation can occur in patients who have been exposed for at least 5 years, keratosis for 14 yr, and cancer for 24.
There was a similar finding for blackfoot disease, which occurred in 8.9 per 1000 population. Histologically, blackfoot disease can be divided into two reaction groups, arteriosclerosis obliterans and thromboangiitis obliterans. The fundamental vascular change in both groups is severe arteriosclerosis leading to arteriosclerotic gangrene in 69.2% (3). Peripheral circulatory disorders occasionally occur in chronic arsenicism (5)(6)(7)(8)(9). At least 15 cases of chronic arsenicism have been reported with gangrene of the extremities (5,7,8,(13)(14)(15)(16)(17)(18)(19). Only two of them were females. The age ranged from 33 to 54. Although detailed pathological observations on autopsy specimens or amputated limbs are lacking in the cases reported, the gangrene appears to result from arteriosclerosis and thromboangiitis obliterans (5). Is arsenic an agent capable of eliciting arteriosclerosis? Hueper (27), in a comprehensive review of the subject of arteriosclerosis, has outlined arsenic as one of the vasodilating depressors. Chronic exposure to arsenical compounds may result in gangrene of extremities which he thought of as the equivalent of focal anoxemic myocardial and cerebral necrosis caused by those agents which exert their main vasodilating effects on the cerebral and coronary arteries.
A dose-response relationship between blackfoot disease and the arsenic level of drinking water was similar to that observed for skin cancer. One final source of evidence to support this positive association is obtained when one examines the frequency of blackfoot disease among young children who were born after the tap water supply was established in 1956. No new cases have been found among those children less than 20 yr old in the area supplied with tap water. However people who changed to tap water could still be affected with blackfoot disease if they had a previous history of drinking artesian well water.
The duration of intake of arsenical water also much influences the severity of permanent impairment of the extremities, mean age at death, and survival rate of patients with blackfoot disease. We found that the association of blackfoot disease with hyperpigmentation, keratosis, or skin cancer was not a coincidental or chance occurrence, which strengthened the likelihood of a causal relationship between blackfoot disease and chronic arsenicism. This leads to the conclusion that skin cancer and blackfoot disease are both part of the entity of chronic arsenicism.