Chronic arsenic toxicity from drinking tubewell water in rural West Bengal

Bulletin of the World Health Organization, 66 (4): 499-506 (1988) © World Health Organization 1988 Chronic arsenic toxicity from drinking tubewell water in rural West Bengal D. N. GUHA MAZUMDER,1 A. K. CHAKRABORTY,2 A. GHOSE,3 J. D. GUPTA,4 D. P. CHAKRABORTY, S. B. DEY,5 & N. CHATTOPADHYAY6 Hepatic damage caused by chronic exposure to arsenic has beenfrequently described. Here we report on 13 patients from West Bengal, India, who consumed large amounts of arsenic in drinking water. An epidemiological investigation of the study area showed evi- dence of chronic arsenical dermatosis and hepatomegaly in 62 (92.5%) of67 members of families who drank contaminated water (arsenic level, 0.2-2 mg/l). In contrast, only six (6.25%) of 96 persons from the same area who drank safe water (arsenic level, < 0.05 mg/l) had non-specific hepatomegaly, while none had skin lesions. Hepatomegaly occurred in all the 13 patients who were studied in detail, althoughfive had splenomegaly. Biopsy of samples of liver from these patients revealed various degrees offibrosis and expansion of the portal zone that resembled non-cirrhotic portalfibrosis (NCPF). In a few villages in West Bengal, India, arsenical dermatosis has been observed in individuals who drink subsoil water that contains a high level of arsenic (1). Among the noncarcinogenic effects of long-term exposure to arsenic are hyperkeratosis, hyperpigmentation, and peripheral vascular and neurological damage (2). Non-cirrhotic portal hyper- tension (NCPH) following chronic arsenic intake has also been reported (3-5), and cirrhosis due to the toxic effects of chronic arsenic exposure has been described (2, 6-8). Non-cirrhotic portal fibrosis (NCPF) is frequently seen in West Bengal (9, 10) and chronic intake of arsenic may produce this condition. We therefore undertook a study to characterize the liver damage in a group of patients who, for prolonged periods, had been drinking water containing high levels of arsenic. An epidemiological study of the patients' area of residence was also carried out to determine whether there was an association betwen hepatic involvement and consumption of drinking water that contained arsenic levels greater than the tolerable limit recom- mended by WHO. ' Professor of Medicine and Chief, Gastroenterology Unit, Department of Medicine, Institute of Post-Graduate Medical Education and Research (I.P.G.M.E. & R.), Calcutta-700 020, India. Requests for reprints should be sent to this author. 2 Professor of Epidemiology, All India Institute of Hygiene and Public Health (A.I.I.H. & P.H.), Calcutta, India. 3 Research Officer, I.P.G.M.E. & R. 4 House Physician, I.P.G.M.E. & R. s Assistant Professor of Sanitary Engineering, A.I.I.H. & P.H. 6 Senior Scientific Officer, Division of Analytical Chemistry, Bhaba Atomic Research Centre, Trombay, Bombay, India. MATERIALS AND METHODS Thirteen patients (8 male, 5 female) from six fam- ilies who lived in two hamlets in the village of Ram- nagar, West Bengal, and who exhibited signs of chronic arsenical dermatosis, were admitted to the Institute of Post-Graduate Medical Education and Research, Calcutta, for investigation. Also, an epi- demiological study of residents of the affected area was carried out by house-to-house visits. Ramnagar is situated 40 km south of Calcutta in the Ganges delta region, far away from industrial areas. The socioeco- nomic status of the people, their occupation, type of housing, dietetic habits, and source of drinking water were recorded. Those surveyed were examined clini- cally, paying particular attention to detection of skin manifestations and liver involvement. Also, water samples were collected from the tubewells used for drinking purposes. The level of arsenic in the water samples was determined spectrophotometrically in the Department of Sanitary Engineering, All India Institute of Hygiene and Public Health (11). The 13 study patients admitted to hospital were thoroughly examined clinically and the following investigations were carried out: routine examination of blood, urine, stools, liver function tests (bilirubin, total protein and albumin, SGOT, SGPT, and alkaline phosphatase levels), determination of prothrombin time, 45 minutes' bromsulphthalein retention test, chest X-ray, and recording of electrocardiogram and electromyogram. All the patients studied were nega- tive for hepatitis B surface antigen (HBsAg). Biopsies were carried out on two samples of liver tissue from 4908 -499- D. N. GUHA MAZUMDER ET AL. each patient. The first sample was stained with haematoxylin-eosin and reticulin, while the arsenic content of the second sample, together with that of samples of hair and nails, was estimated by neutron activation analysis (12). Upper gastrointestinal endo- scopy and splenoportal venography were performed and the intrasplenic pressure was measured in order to detect any oesophageal varices or portal hyper- tension (13). Unfortunately, no haemodynamic studies could be performed since the patients refused permission. RESULTS The results ofthe epidemiological survey ofthe two affected hamlets ofRamnagar showed that most of the inhabitants were of a low socioeconomic class and lived in mud houses; only two of the families resided in brick hduses. The source of drinking water for all the people in the village was a shallow tubewell (depth, 36.5-45.7 m). Most of the inhabitants worked the land or were day labourers, though a few belonged to middle-class families and were employed in offices. None of the people habitually ate seafood. A group of 48 persons from eight families who lived in one cluster drank water from a public tubewell (number 1) that contained a very high level of arsenic (2 mg/l) (see Table 1); in addition to mottled hyperpigmentation of the skin, 46 (95.8%) of these persons had hepatomegaly. The two members of the families who were not affected had lived elsewhere for some time. Also, of two unaffected members of the family that drank water from tubeweil number 2, one was a baby aged 4 months and the other was a woman who had come to the village only a few months previously after her marriage. The one unaffected member of the family that drank water from tubewell number 3 was a baby aged 6 months. Although the overall prevalence ofhepatomegaly was 92.6% (62 out of 67) among people drinking water that contained levels of arsenic considerably greater than the limit recommended by WHO (0.05 mg/l) (14), the prevalence was much lower (6 out of 96; 6.25 %) among residents of the same -area who drank water containing a level of arsenic below this limit. The ages of the 13 patients investigated in detail ranged from 15 to 50 years (10 were aged 20-40 Table 1. Epidemiological data for Ramnagar village, West Bengal Arsenic No. of people No. of No. of Village hamlet and concentration of drinking people with people with Attack rate of tubewell number tubewell water (mg/I) tubewell water skin pigmentation hepatomegaly hepatomegaly (%) Arsenic concentration > 0.05 mg/l 1 2.0 48 46 46 95.8 Gharamipara 2 0.2 8 6 6 75 3 0.725 7 6 6 85 Bakerpara 4 0.2 4 4 4 100 Total 67 62 62 92.5 Arsenic concentration < 0.05 mg/l 5 ND" 11 0 0 Gharamipara 6 0.035 8 0 0 7 ND 26 0 3 8 0.037 16 0 1 i9 0.022 7 0 1 Bakerpara 10 ND 12 0 0 (1 1 0.031 16 0 1 Total 96 0 6 6.25 ND - not detectable. 500 Fig 1. Photomicrograph of liver biopsy from a patient showing expansion of portal zone containing a few round cells (case 5) (haematoxylin-eosin stain, x200). (a) (b) (C) (d) Fig 2. Photomicrographs of liver biopsy specimens from study patients. (a) Grade I fibrosis-mild fibrosis producing expansion of portal zone (case 8). (b) Grade II fibrosis-expansion of portal zone with thin fibrous extension producing septae (case 6). (c) Grade Ill fibrosis-moderate fibrosis in the portal zone with thick septae; tendency to mild pseudolobulation is seen with slender bands. (d) Grade IV fibrosis-dense fibrosis within the liver with a marked tendency to pseudolobule formation (case 4) (reticulin stain x 25). Fig 3. Photomicrograph of biopsy from a patient showing a portal zone with multiple vascular channels and significant fibrosis (case 10) (haematoxylin- eosin stain, x 200). CHRONIC ARSENIC TOXICITY IN RURAL WEST BENGAL years). The clinical features presented by these patients are given in Table 2. All exhibited mottled dark-brown (rain drop) pigmentation distributed over the whole body and limbs, while 10 patients had hyperkeratosis of the palms and soles. Also, in six patients the pigmentation was present on the mucous membrane of the oral cavity, including the underside of the tongue. The liver was palpable and firm in all patients. Gross splenomegaly (7 cm) occurred in only one patient, who also had ascites and exhibited haematemesis and melena on two occasions during her stay in the hospital (1.5 months). The spleens of four other patients were also palpable ( < 3 cm). The haemoglobin levels in the 13 patients varied from 10 g to 13 g per 100 ml blood. No abnormality was detected in the differential lymphocyte count or in the blood sugar, urea, and creatinine levels of any of the patients. Traces of albumin were found in the urine of two patients. Also, Ascaris lumbricoides was detected in the stools of six patients, Entamoeba histolytica in those of three patients, while Ancylostoma duodenale and Trichuris trichiura were found in the stools of one patient each. Although the results of routine liver function tests were normal, the prothrombin time for one patient was elevated, while the bromsulphthalein retention tests ofthree patients were abnormal (13.8%, 13.1 %, and 15.3%). The electrocardiograms of all patients were normal, but the electromyograms of the eight Table 2. Clinical features among the 13 patients investi- gated in the study Clinical feature No. of patients Anorexia 4 Nausea 3 Vomiting 3 Diarrhoea 1 Abdominal pain 10 Heartburn 2 Tingling and numbness 10 of hands and feet Pigmentation 13 Thickening of palms 10 and soles Anaemia 13 Weakness of limbs 2 Hepatomegaly (2-4 cm) 13 Splenomegaly (3-7 cm) 5 Ascites 1 patients from whom these measurements were ob- tained showed mild to moderate neuropathic patterns. Histological studies of samples of liver showed en- largement and fibrosis of varying degrees in the portal regions (Fig. 1 and 2). Fibrous extensions from the portal tracts that produced septae were observed in 10 patients, six of whom showed a tendency to pseudolobule formation. The limiting plates ofparen- chyma were intact in most patients, but in some, advancing fibrous tissue (septae) divided the liver lobules into small segments. Slight-to-moderate ac- cumulations of lymphoid cells were visible in the portal tracts and in the interlobular areas; however, frank cirrhosis was not observed in any case. In some patients, portal vein branches were replaced by mul- tiple vascular channels (Fig. 3). Bile ductular pro- liferation occurred also in the dense fibrous portal tracts and septae. However, gross abnormalities were not observed in the liver parenchyma and there was no accumulation of cells in the sinusoids. A liver biopsy of the only patient who had ascites, haema- temesis, melena, and gross splenomegaly (case 1) yielded only scar tissue and fragmented liver lobule. Unfortunately, this patient did not consent to a further biopsy or to other investigations except endoscopy. Oesophageal varices were observed in this patient and in two of the other four patients in whom spleno- megaly was identified by upper gastrointestinal en- doscopy. Splenoportal venography indicated that four of the five patients with splenomegaly had increased intrasplenic pressure (30-36 cm saline) and evidence of intrahepatic portal vein obstruction. The intra- splenic pressure in the other nine patients studied was normal (16-20 cm saline). The level of arsenic in the tubewell water drunk by the patients, together with the duration of their intake and the level of arsenic in patients' hair, nails, and liver biopsy material are shown in Table 3. The level of arsenic in the water was much higher than the value of 0.05 mg/l recommended by WHO as the limiting concentration, while its level in hair, nails, and liver tissues was also greater than the levels in the controls. DISCUSSION The results of the study indicate chronic toxicity caused by drinking tubewell water that contained a high level of arsenic in a village in the Ganges delta region far away from sources of industrial pollution. Generally, the total arsenic level of ground and sur- face water in this area is less than 10 .g/l, although levels greater than 1 mg/l have been recorded for deep tubewell water or in areas where industrial efflu- ent has led to severe contamination (2). It is therefore interesting that the village in our study used water 503 D. N. GUHA MAZUMDER ET AL. Table 3. Relationship between the amount of arsenic ingested by patients, its level in various body tissues, and the severity of hepatic lesions Arsenic level (mg/kg) in: Liver lesionsb: Arsenic con- Duration Multiple Age centration in of water Fibrosis Enlargement Tendency to vascular (years)/ tubewell intake Liver in portal of portal pseudolobule channel in Patient sex' water (mg/I) (years) Hair Nails (dry weight) zone zone formation portal tract 1 43/F 0.75 lOC 7.5 24 NDd + + + +' 2 26/F 2 4 13.3 66 1.5 + none none + 3 25/F 2 4 1.4 40 1 + + + none + 4 23/F 2 4 7.1 57 0.5 + + + + + + + 5 47/M 2 1 c 7.7 27 3 + + none + 6 22/M 2 1 3.8 16 1.5 + + + + + 7 36/M 2 4c 20 41 ND + + + none + 8 31 /M 2 4c 7.4 31 1.5 + + + + 9 30/M 2 4 ND 24 1 + + + + + + + 10 30/M 0.5 20C 4.7 26 ND + + + + + + + 11 25/M 0.22 1 7.1 ND 2 + none none none 12 40/F 2 1 6 29 6 + + + + + + 13 18/M 2 1 3.6 27 3 + + + + none + Control values < 0.05 0.15±0.35 0.34±0.25 0.16±0.04 (n= 96) (n= 5); (n- 5); (n-5); P<0.001 P< 0.001 P<0.01 M-male; F-female. b Grading of fibrosis: + to + + + +. c Presence of portal hypertension. d ND - not detectable. Unsatisfactory liver biopsy showing scar tissue and fragmented liver lobule. The patient had ascites, gross splenomegaly, and a history of haematemesis. from shallow tubewells located far away from in- dustrial activity. Arsenic could not be detected in samples of water from deep tubewells in the area. None of the affected people was in the habit of eating seafood, which is known to contain a high level of arsenic (2). Similarly, samples of soil from the area did not contain any appreciable levels of arsenic (15). Thus, although the source of the arsenic was water from shallow tubewells, its origin was not deter- mined. The results of the epidemiological study showed not only evidence of the typical mottled skin pigmentation caused by arsenic toxicity but also of hepatomegaly in most of those who drank water con- taining arsenic at levels greater than the 0.05 mg/l limiting concentration recommended by WHO (14). Hepatomegaly was observed in 92.5% of those (67) who drank the contaminated water (arsenic content 0.2-2 mg/l), while six of the 96 individuals (6.25%) who drank water containing less than 0.5 mg/l arsenic had mild non-specific hepatomegaly but no skin changes. All of the 13 patients with arsenic toxicity who were investigated in the study had skin pigmentation and hepatomegaly. Evidence of neuropathy was found in the eight patients investigated by electro- myography, although, in contrast to previous reports, none showed any ECG abnormality or sign of per- ipheral vascular disease (2). Although liver damage caused by chronic arsenic toxicity has been reported frequently, its nature and effect on portal haemodynamics have been variously described (2-8, 16-18). We have presented here data on the largest reported series of patients who exhibi- ted portal fibrosis following intake of drinking water contaminated with arsenic. However, definite evi- dence of portal hypertension was observed in only five cases. The almost normal liver function tests and absence of gross damage to hepatocytes in the study 504 CHRONIC ARSENIC TOXICITY IN RURAL WEST BENGAL cases are consistent with previous reports (3, 5). Also, the features of the periportal fibrosis and mul- tiple vascular channels in expanded portal zones are in accord with the results of other studies of NCPF in India (5, 9, 19-21). The hepatic histology described by Morris et al. (3) in cases of NCPH caused by chronic arsenic toxicity did not include extension of fibrosis from the portal tract or any tendency to the pseudolobule formation that we saw in our cases. However, like us, they did observe an expansion of the portal zone. In contrast, Franklin et al. (7) repor- ted periportal fibrosis that extended intralobularly in one of the three cases of arsenic toxicity they studied. Fibrosis of the portal tract with septae in two patients was also described by Rosenberg (8). Finally, Datta et al. (5) reported a higher level of arsenic in the livers of four out of nine patients who had NCPF with portal hypertension, although only two of them had drunk water containing a high concentration of arsenic. Also in the last-mentioned study, skin mani- festations, such as typical pigmentation of the body or hyperkeratosis of the palms and soles, occurred in only one patient. The arsenic level in the hair of five normal patients in our study was 0.15 ± 0.35 mg/kg, which is con- sistent with previously published values (2). Although it has been established that chronic arsenic intake can produce NCPF, the results of the present study indicate that portal hypertension is not a necessary consequence. The absence of any precise correlation between the level of arsenic in liver tissue and portal venous pressure has been reported pre- viously (5) as has the lack of clear correlation be- tween the degree of portal fibrosis and the magnitude of portal pressure (21). The etiology of portal hyper- tension in patients with chronic arsenic toxicity has been ascribed to changes in the branches of veins in the portal tracts (3, 5), which is consistent with the vascular changes that we observed in patients with portal hypertension. However, we did not observe vascular lesions such as intimal thickening of arteries in the portal region or the cirrhotic changes described by Rosenberg (8) in children who drank arsenic- contaminated water. Although we examined only samples of liver obtained by needle biopsy, such changes have not been reported for wedge biopsy studies of liver on patients with chronic arsenic toxicity (3, 5). ACKNOWLEDGEMENTS The authors are grateful to Dr S. K. Gupta, Dr K. P. Sen Gupta, Dr B. N. Guha Ray, Dr M. Sankar Das, Dr P. K. Bhattacharjee, and Dr S. Bhattacharjee for their help in carrying out this work. RESUME TOXICITE ARSENICALE CHRONIQUE DE L'EAU DE BOISSON PROVENANT DE PUITS INSTANTANtS DANS LES RtGIONS RURALES DU BENGALE-OCCIDENTAL On observe dans quelques villages du Bengale-Occi- dental, en Inde, une dermatose arsenicale chez les sujets qui boivent de l'eau contenant de fortes concentrations d'arsenic. Par le passd, on a signale diverses atteintes hepatiques chez des sujets souffrant d'une intoxication arsenicale chronique. Sont ici decrits les rdsultats d'une etude approfondie portant sur 13 malades atteints d'intoxi- cation arsenicale chronique, etude au cours de laquelle on s'est particulierement attache a etablir la nature des atteintes hepatiques. Une etude epidemiologique de la region a indique que 62 personnes sur 67 (92,5 %) buvaient de l'eau contaminee par de l'arsenic (concentration d'arsenic: 0,2-2 mg/l) et mon- traient tous les signes d'une dermatose arsenicale chronique et d'une hepatomegalie, tandis que 6 personnes sur les 96 etudiees qui ne buvaient pas d'eau contaminee (6,25%) (concentration d'arsenic inf6rieure a 0,05 mg/l) ne mon- traient qu'une hepatomegalie non specifique et ne presen- taient aucune manifestation cutanee. Une hepatomegalie d'importance variable (2 a 4 cm) a ete observee chez les 13 sujets de l'etude approfondie. On a 6galement observe une splenomegalie (3 a 7 cm) chez 5 de ces malades. Les epreuves classiques de la fonction hepa- tique se sont revelees normales chez tous ces malades, mais le temps de Quick etait anormal chez l'un d'eux, et 3 autres montraient des anomalies dans l'epreuve a la brome- sulfonephtaleine. Tous les electrocardiogrammes etaient normaux, mais les electromyogrammes des 8 sujets chez qui on les a obtenus ont montre des lesions nerveuses benignes a mod6rees. On a observe dans le foie de tous les sujets etudies une hypertrophie des espaces portes avec divers degres de fibrose. On a egalement observe des extensions fibreuses a partir des espaces portes, avec apparition de septums chez 10 malades, dont 6 montraient une tendance a 505 506 D. N. GUHA MAZUMDER ET AL. la pseudolobulation. Aucune cirrhose franche n'a ete observ6e. Chez un des malades, les troncs veineux avaient et6 remplaces par des derivations vasculaires multiples dans la zone porte. En outre, 5 malades ont montre une hypertension portale, mise en evidence par une pression intrasplenique 6levee (30 a 36 cm de solut6 physiologique). L'endoscopie a revele des varices oesophagiennes chez 3 malades. On a d6termine la concentration en arsenic de preleve- ments de cheveux, d'ongles et de tissu hepatique chez tous les malades et elle s'est averee plus elevee que chez les t6moins. Cette etude a mis en lumiere le fait que l'intoxi- cation chronique par l'arsenic est relativement commune en Inde chez les personnes qui boivent de l'eau contaminee puis6e dans le sous-sol. L'atteinte h6patique est un sympt6me d'accompagnement important de cette intoxi- cation et les lesions observees sont identiques a celles qui accompagnent une fibrose portale non cirrhotique. REFERENCES 1. GoRAu, R. ET AL. Chronic arsenic poisoning from tube- well water. Journal ofthe Indian Medical Association, 82: 34-35 (1984). 2. Arsenic. Geneva, World Health Organization, 1981 (Environmental Health Criteria Series, No. 18). 3. MORRIS, J. S. ET AL. Arsenic and non-cirrhotic portal hypertension. Gastroenterology, 64: 86-94 (1974). 4. DATTA, D. V. Arsenic and non-cirrhotic portal hyper- tension. Lancet, 1: 433 (1976). 5. DATTA, D. V. ET AL. Chronic oral arsenic intoxication as a possible factor in idiopathic portal hypertension (non-cirrhotic portal fibrosis) in India. Gut, 20: 378-384 (1979). 6. HUTCHINSON, J. Diet and therapeutics. Archives ofsur- gery, 6: 389-391 (1895). 7. FRANKLIN, M. ET AL. Fowler's solution as an etiologic agent in cirrhosis. American journal of the medical sciences, 219: 589-596 (1950). 8. ROSENBERG, H. G. Systemic arterial disease and chronic arsenicism in infants. Archives ofpathology, 96: 360-365 (1974). 9. BASU, A. K. ET AL. Non-cirrhotic portal fibrosis with portal hypertension-a new syndrome. Parts I and II. Indian journal of medical research, 55: 336-359 (1968). 10. GUHA MAZUMDER, D. N. ET AL. Search for aetiologi- cal factors of non-cirrhotic portal fibrosis. Indianjour- nal ofgastroenterology, 3: 25-26 (1984). 11. Standard methods for examination of water and waste water, 14th edition. New York, American Public Health Association, 1967, p. 283. 12. OBRUSNIK, I. ET AL. Indication of environmental pol- lution by means ofINNA ofthe hair ofsome free-living mammals. Journal of radioanalytical and nuclear chemistry, 83: 397-406 (1984). 13. GUHA RAY, B. N. ET AL. Direct splenocaval shunt for selective decompression of portal hypertension in chil- dren. Surgery, 87: 271-279 (1980). 14. Guidelinesfor drinking water quality, Volume 1. World Health Organization, Geneva, 1984. 15. CHAKRABORTY, A. K. ET AL. Arsenical dermatosis from tubewell water in West Bengal. Indian journal of medical research, 85: 326-334 (1987). 16. CLENCOPATHOLOGICAL CONFERENCE. Chronic arseni- cal poisoning and non-cirrhotic portal hypertension-a case for diagnosis. British medical journal, 4: 725 (1971). 17. HUET, P. M. ET AL. Non-cirrhotic presinusoidal portal hypertension associated with chronic arsenical intoxi- cation. Gastroenterology, 68: 1270-1277 (1975). 18. VIALLET, A. ET AL. Presinusoidal portal hypertension following chronic arsenic administration. Gastroen- terology, 62: 177 (1972). 19. Workshop on Non-Cirrhotic Portal Fibrosis. New Delhi, Indian Council of Medical Research, 1969, pp. 36-45. 20. SAMA, S. K. ET AL. Non-cirrhotic portal fibrosis. American journal ofmedicine, 51: 160-169 (1971). 21. NAYAK, N. C. ET AL. Obliterative portal venopathy of the liver associated with so-called idiopathic portal hypertension or tropical splenomegaly. Archives of pathology, 87: 359-369 (1969).