Dawit Abebe, Asfaw Debella, Frew Tekabe, Yared Mekonnen, Ambaye Degefa, Asnakech Mekonnen, Alemtsehaye Tefera, Hailu

Mamo, Bekure Tsegaye. 2011.Ethiop Med J, Vol. 49, Supplement 1

AN OUTBREAK OF LIVER DISEASE IN TAHTAY KORARO WOREDA, TIGRAY REGION OF ETHIOPIA: A CASE-CONTROL STUDY FOR THE IDENTIFICATION OF THE ETIOLOGIC AGENT
Dawit Abebe, 1 Asfaw Debella, 1 Frew Tekabe,1Yared Mekonnen, 1 Ambaye Degefa, 1Asnakech Mekonnen, 1 2 Alemtsehaye Tefera, Hailu Mamo,1 Bekure Tsegaye

ABSTRACT
Background: An outbreak of fatal liver disease of unidentified origin in Tahtay Koraro Woreda, Tigray Region was reported at the end of 2005. In response to this report, a team of investigators was deployed from the EHNRI to assess the situation in the affected area. Objectives: The objective of the trip was to investigate the probable etiological agent(s) for the stated health problem and to suggest possible means of containing the outbreak. Method: A case-control epidemiological method was employed in collecting information from the affected (case) area and a non-affected adjacent area (control) using structured questionnaire. Qualitative data was also collected through focus group discussion (FGD) at community as well as woreda level. Samples of biological and environmental nature were also collected from both case and control areas for a subsequent laboratory analysis in order to identify the causative agent of the outbreak. Result: The comparative analysis of the various variables and possible exposure factors between the two sites (case vs control village) revealed that both villages were similar in terms of their exposure to most of the suspected factors including exposure to chemicals, food storage and handling practices, and to the use of traditional herbal medicines or consumption of wild edible plants. However, it was observed that the residents in the affected site (over 96%) relied more on unprotected or protected well as a source of drinking water and other household purposes while most in the non-affected depended on fresh water from river or unprotected spring. This difference was significant, p=0.000 (OR = 840, 95% CI = 50.1, 14095.7). Conclusion: Though several of the possible exposure factors to the epidemic appear to be similar for both the case and control villages, the water source they use for drinking and other household purposes differed markedly. It therefore seems plausible that the problem in the case area could be linked to this water source. Key words: Liver disease, Case-control, FGD indigenous plant remedies and local alcoholic beverages were implicated for 75% of men with cirrhotic conditions (3). Hepatitis-B virus and Schistosoma mansoni infection were also considered as the probable etiologic agents for the high prevalence of mixed types of cirrhosis and other forms of liver diseases in Ethiopian patients though there is no mention of veno-occlusive liver disorder in the country (4, 5). Hepatic veno-occlusive liver disease (VOD) in humans is often caused by pyrrolizidine alkaloid (PA) poisoning and has usually been misdiagnosed as viral

INTRODUCTION
A number of morbidity and mortality data indicate that hepatitis, cirrhosis of the liver and hepatoma are prevalent in Ethiopia. For example, in a study conducted some decades ago, about 11.5% of deaths on medical wards were linked to liver disease (1). 7.5 per cent cirrhosis of the liver was reported in 1562 consecutive autopsies conducted at Menelik II Hospital of Addis Ababa (2). History of regular use of
1 2

Ethiopian Health and Nutrition Research Institute, Addis Ababa, Ethiopia, P. O. Box 1242, Drug Administration and Control Authority, Quality Control and Toxicology Laboratory, Addis Ababa, Ethiopia.

hepatitis or cirrhosis. Several environmental hepatotoxins are ubiquitous in nature. Pyrrolizidine alkaloids found in herbal medical preparations and bush teas are the major recognized causes of VOD. Other causes of hepatic veno-occlusive disease include radiation treatments to the liver, bone marrow transplantation, chemotherapeutic drugs, birth control pills, immunosuppressive therapy, and intravenous administration of vitamin-E preparation to low birth weight infants (6). It is not easy to determine duration and toxic dose of PAs since this depends on the type of the alkaloids which vary widely among themselves in potential toxicity, on the part of the plant, on the season the plants are to be found, on the age and physical stamina of the individuals exposed to these alkaloids. However, it has been suggested by the World Health organization in 1989 that the lowest intake rate of PAs that reportedly caused VOD in humans was just 0.015mg/kg of body weight per day. For a 70kg adult, that would correspond to 1mg total per day (7). Although the main site of toxic reaction to the PAs is the liver, a few PAs act on other organ systems or the nervous system. This is assumed to be due to their long half-life, allowing them to migrate to other parts of the body prior to further metabolism (8). A report based on epidemiological and clinical evidence with a significant morbidity and mortality in Tahtay Koraro Woreda in Northern Tigray gathered by a team of experts from the Faculty of Medicine, Addis Ababa University, was communicated to the Ethiopian Health and Nutrition Research Institute (EHNRI) in the first week of December 2005. Along with the report, samples of consumable items collected from the affected area were also submitted for laboratory analysis (9). On the basis of the report of the team of experts from the Faculty of Medicine, Addis Ababa University, the preliminary laboratory findings on the submitted samples and the fact that the alleged health problem is of localized nature without affecting the nearby or adjacent village(s), a working hypothesis was postulated that the outbreak could be more likely of toxic rather than of infectious origin. Against this backdrop of suggestive veno-occlusive liver disease among the community members of Tsaeda Emba village in Tahtay Koraro woreda, the need for making enquiries about possible risk factors that predispose patients to this condition has become evident. A multi-disciplinary task force from the EHNRI was therefore dispatched in December, 2005 to

investigate environmental and xenobiotic risk factors (hepatitis B, C, use of herbal remedies, consumption of wild edible plants, application of pesticides, insecticides, herbicides, pollutants of drinking water sources, storage and preparation of food items, etc.) in order to formulate intervention strategy for the prompt discontinuation or removal of the offending toxin and/or suggest treatment regimen (if any).

MATERIALS AND METHODS

Study site: A rapid field survey assessment was conducted in mid of December 2005 in the affected village, Tsaeda Emba and the nonaffected village, Jek Emba, both in Tahtay Koraro woreda, Study design: In such outbreak of a disease with unspecified origin, utilization of rapid epidemiological method is crucial to rule out or rule in the possible causes of the problem (10, 11). A case-control design was, therefore, employed (Table 1). The affected village, Tsaeda Emba served as a case and the adjacent village, Jek Emba, was used as a control. Lab. analysis of the samples:-Organoleptic tests (examination of color, odor and texture), qualitative chemical tests on the presence of aflatoxins, pyrollizidine alkaloids and pesticides particularly organochlorines was carried out for the five samples namely, sample no. 1 flour of an unidentified grain, sample no.2 mixture of seeds of millet, sorghum and maize, sample no.3 Teff admixed with straw, sample no.4 mixture of maize, powdered grain (no label was shown on the type of grain), straw, soil and rat excreta, sample no. 5 maize using standard methods (12, 13). In addition to this analysis of water sample (sample no.6) for Copper (Cu), Arsenic (As) and cyaogenic glycoside were done using official methods (14, 15). Furthermore, feeding the grain samples to laboratory mice to simulate the toxicity occurred in humans and identify possible intoxicants was also carried out as described by Loomis, 1978 (16). . Data collection and sampling techniques: A house to house interview was conducted in 30 households each in case and control that were two different adjacent villages having the same agro-ecological and socio-economic conditions using structured questionnaire. The interview was made with the household heads. Health workers (nurses) were used as data collectors following proper orientation by the investi-

gating team on how to fill the questionnaires. Focus group discussion was made at the community level in the affected locality involving about twenty each of the women and men participants separately. Similar discussions were held at the woreda level with experts and representatives of the health office, education, agricultural and rural development, and the medical director of Shire Hospital. Information was also obtained through key informants such as elders, health and agricultural extension workers and teachers who are knowledgeable about the woreda. Most of the questions posed to the participants in the various groups revolved around: drinking water sources, exposure to chemicals, food storage and handling process, the use of traditional herbal medicines and wild edible plants, socio-economic, demographic, etc. characteristics. Samples of stored grain and flour were collected for tests of heavy metals, organophosphorous compounds, toxic plant secondary metabolites and aflatoxins. Samples of consumables such as edible oil, cooked foods and Tela (locally made alcoholic beverage) were collected from the residents in the affected and non-affected villages for laboratory analysis. Wild edible plants and parts of plants used for various household purposes were also collected from residents in the affected village. Samples of water from drinking water sources, i.e., river/spring, protected and unprotected well were also collected both from the case and control sites. Samples of plants found growing fully or partially submerged in the most commonly used water source, i.e., the unprotected well in the affected village were gathered for scientific identification and subsequent laboratory analysis. Blood samples were also collected from persons with self-diagnosed ill-health in the case area. The collected samples were stored and transported to Addis Ababa using cool box containing icepack. All the collected samples were used for various analysis. The methodology employed for the laboratory analysis of the collected samples were described elsewhere in this edition of the Journal. Data analysis: Data were entered using SPSS and analysis was performed using STATA software. The statistical analysis was limited to univariate comparison of possible exposure factors between the two groups.

RESULTS

Laboratory analysis of samples: About six samples of consumable items previously collected from the affected area by the expert group of A.A.U. and submitted to our Institute were subjected to laboratory analysis even though most of them were inadequate, unrepresentative or poorly labeled. Thus, of the six samples submitted, the one which was coded as sample No.4 (referred to a sample containing mixture of maize, unidentified powdered grain, straw, soil and rat excreta) was found to be contaminated by pyrrolizidine alkaloids and aflatoxins. Feeding trials of this sample in laboratory animals also showed to have anorexic effect for the first 24-hour period which was followed by the death of all the animals after the 4th day of feeding. Histopathological examination of the liver of the dead animals also revealed severe hepatic necrosis. Though the rest of the samples are free from possible intoxicants at preliminary screening stage, it was difficult to make appropriate conclusion owing to the inadequacy and unrepresentative nature of the commodities. Findings of the laboratory analysis of the various samples that were collected by the investigating team of EHNRI are discussed elsewhere in this edition of the Journal. Field survey result: Topography and general observations: The village of Tsaeda Emba is 1500m above sea level and is situated on a plateau surrounded by chain of mountains that are sparsely covered with shrubbery. The ecology is highly degraded, with only scattered shrubs and trees and practically with no ground cover. The soil is grayish sandy-loam and appears to be exhausted as can be deduced from the abundance of pebbles, rocks and boulders. Months of major precipitation are June-August. Average annual rainfall is said to be 800-900mm and hailstorm is the major natural calamity to agricultural production rather than recurrent drought. The affected locality is generally self-sufficient in food production and, therefore, external food aid is a rare phenomenon. Average daily temperature ranges from 25 to 270C with considerably low rate of humidity.

Table 1. Instrumental tool used for rapid investigation of the out break the liver disease during the field survey.

Study design Case-control design Within the affected community

Purpose -To assesses and compare possible HH as well as individual level exposures between affected & unaffected households Possible exposure categories at HH/individual level: 1. Environmental sanitation (water) 2. Exposure to toxic chemicals 3. Food storage and handling 4. Exposure to wild edible plants/traditional medicine 5 Socio-economic/demographic/occupation expo sures 6. Other exposures(infected animal products, drug use, etc) -To assesses and compare possible Wereda/ community level exposures between affected & unaffected communities Possible exposure categories at Wereda level: 1. Environmental sanitation (water) 2. Exposure to toxic chemicals 3. Food storage and handling 4. Exposure to wild edible plants/traditional medicine 5. Socio-economic/demographic/occupation exposures 6. Other exposures(infected animal products, drug use, etc) - To understand and describe the following: - Who is affected? Age, sex, SOS, etc - Is there any clustering of the disease? - How did it occur? -When did it occur? -What are the symptoms? -Possible diagnosis? -Any Treatment? -Description of the Morbidity pattern? -Mortality patter? -To document community awareness and perception of the outbreak, its possible causes, health seeking behaviors, available intervention to contain the outbreak

Method of data collection/ analysis - This was done through a rapid house-to-house survey - Sample size: 30 HHs in affected & 30 HHs in the unaffected villages - Data was collected using standard questionnaire from both affected & unaffected households - Comparative analysis b/n HHs of affected & unaffected vil lages was done - This was done through keyinformants interview using sem - structured questionnaires - Zone health office, agricultural office, extension agents, Wereda health bureau, agricultural bureau, in affected & unaffected neighboring Wereda were interviwed - ecological/correlation assess ment

Ecological/ Correlation comparison of affected Wereda with unaffected nearby Wereda

Descriptive study asic epidemiological description of the outbreak)

- Key informant interview with health workers - Key informant interview with community health agents/ extension health workers - Health service statistics/records about the outbreak: No. of patients, fatality, treatment, etc - Key informant interview with community members/teachers - Focus group discussion with community members (at least 2) (1 with female & 1 with men) -Informal discussion with Wereda health bureau ( field survey team will do the collection)

Descriptive study (Community awareness & perception of the outbreak, any activities, etc?) Collection of samples

- Search for causative agents: Edible wild Plant, Stored food, Water sample, oil samples, local liquor beverages, cooked and uncooked staple foods etc

Staple foods and dietary habits: Major agricultural crops are millet, sorghum and maize, the latter two constituting the most important part of the staple diet. Millet is mainly used in brewing Tela. Like most other places in Ethiopia, Injera (pancake-like thin bread) is consumed with wot (stew or sauce) or red pepper in both sites that were studied. Except for chilly peppers, other vegetables or root crops are extremely uncommon. However, fruits of certain wild plants such as Vangueria edulis (guramayile), Diospyros mespiliformis (aye), Ximenia americana (enkoy) are commonly consumed when available. Though there are only a few pack animals (mainly donkeys), cattle, goats and chicken are common. Consumption of Tela is common in the affected village. Focus group discussion: Separate FGDs were conducted with 20 men and women residents in each group of Tsaeda Emba (case village). All the participants acknowledged the occurrence of the disease in the community and that they commonly refer to it as Ebete (Tigringa), which literally means swelling. They stated that the disease has been in the community for about 4-5 years affecting mainly those between the ages of 2-30 years. They all claim that they have no definite idea as to what causes the disease though initially they thought it could be from the water of the open well they use along with other wild beasts. Nevertheless, with the erection of a hand pump beginning from September, 2004, reliance on the open well has decreased or used only when there is a breakdown of the hand pump which the participants say is a common phenomenon. Despite such intervention measure the disease, however, continued unabated, according to most of the participants. Thus they rule out the water source as possible cause of the scourge as they also assert that even a few of those who drink from stream or river are being struck by the disease. The participants allege that 80 people had so far died of the disease in the village and currently about six patients are being in a poor state of health in addition to many others that are hospitalized in Mekele and Shire towns. The pick season of the epidemic, according to them, is in the months of July and August. The major symptoms as characterized by the participants of the discussion include abdominal dropsy, loss of weight, change in skin color, crump, vomiting, eye color turning greenish (jaundice), leg swell-

ing, and fatigue. Febrile condition is excluded from the list of the symptoms. Patients with the disease are said to have a good appetite and, therefore, those who can afford to eat frequently last longer than those that are less fortunate to do so. They observe that no patient has ever been recovered so far from the disease or resumed productive economic activity. They dont believe that the disease is communicable and there is no alternative traditional therapeutic method prescribed for treatment. The participants claimed that the disease has also affected their livestock though the symptoms they described in this case do not correspond to those in humans. They expressed that they are very anxious of the disease and, therefore, they want to flee the village or be resettled elsewhere. The investigating team also noted that the use of commercial fertilizers is uncommon. The community rather heavily relies on organic manure for agricultural production. Herbicides are rarely used even though the exotic weed referred to as Parthinum sp. (kinchie, Tigrigna) poses considerable nuisance. The use of other pesticides is also uncommon or if there is any need for application of these, it is according to the guidance and strict follow up of the agricultural development agents. Counter to the claims of some FGD participants in the affected village, officials of the Woreda Agricultural office refute the use of DDT as pest control strategy on harvested crops. The team also realized that some of the women make use of melted polyethylene plastic bags and castor seeds (residue of which contains the toxalbumin, ricin) and seeds of Argemone mexicana (talayan dender) to grease Injera baking pan. The seeds of the latter species contain sanguinarine alkaloids (that caused an outbreak of epidemic dropsy in India) are used as potentiating ingredient of the local alcoholic beverage, Tela. Participants of the FGD and/or key informant interview at Woreda level in Shire town stated that two people in a different Woreda known as Asgedet, have shown similar symptoms like those in Tahtay Koraro though they were not sure that these patients arrived recently or long time residents. The symptoms, ascites, abdominal crump, bloody diarrhea, nasal bleeding, and irregular fever more or less overlap with those identified by the FGD participants of the community in the affected village. The participants of the FGD at Woreda level also confirmed the claim of the group interview at community level regarding good appetite of the patients suffering from the disease. They estimate that patients

Table 2. Assessment of comparative exposure factors to liver disease of an unidentified origin

No. of respondents in liver

disease affected (case) and non- affected (control) area Exposure Category 1. Drinking water source Variable Exposed (Unprotected open &/or protected well) Not exposed (Fresh water of river &/or unprotected spring) Exposed (Yes) Not exposed (Never) Exposed (Yes) Not exposed (Never) Exposed (Yes) Not exposed (Never) 5. Use of chemical fertilizers Exposed (Yes) Not exposed (Never) Exposed (Yes) Not exposed (Never) Exposed (Yes) Not exposed (Never) 27 (90) 3 (10) 23 (76.7) 7 (23.3) 0.19 (0.03,1.0) 2.7 (0.6,11.8) N.S. 29 (96.7) 1 (3.3) 840 (50.1,1409 5.7) 0.000
Cases (%) Controls (%)

OR (95% CI)

P-value

1 (3.3)

29 (96.7)

2. Use of DDT in and around water sources

2 (6.7) 28 (93.3) 8 (26.7)

2 (6.7) 28 (93.3) 21 (70) 0.15 (0.05,0.48) 1 (0.1,7.6) N.S.

3. Use of pesticides on stored grains

22 (73.3) 5 (16.7)

9 (30) 0 (0)

0.001

4. Use of herbicides or pesticides on field crops

25 (83.3)

30 (100)

4379 (0.0,4.4)

N.S.

6. Storage of grain for 6 months

22 (73.3) 8 (26.7)

28 (93.3) 2 (6.7)

N.S.

7. Use of traditional medicinal plants

7 (23.3)

2 (6.7) 4.2 (0.8,22.5)

23 (76.7)

28 (93.3)

N.S.

DISCUSSION
Preliminary report of the Faculty of Medicine, Addis Ababa University, and the laboratory investigation conducted on one of the samples submitted along with this report were suggestive of liver VOD. This coupled with the nature of the affliction and its localized occurrence in the affected village for the last 3-5 years appears to indicate that the disease might be a sub-chronic/chronic disorder that could have possibly originated from exposure to toxic substance (s) rather than due to an infectious agent. Furthermore, the major symptoms of the disease as characterized by the participants of the FGD at the community and woreda level and the alleged poor prognosis could well be associated with liver abnormality stemming from toxic substance (s). However, during the present investigation a link between the stated health problem and exposure to potential environmental toxins such as pesticides/ herbicides on field crops and commercial fertilizers is safely ruled out owing to none use or sporadic use of these chemical agents in the affected village and to the statistically insignificant differences in use between the two groups of respondents. Significantly higher use of pesticides on stored crops in the control area than in the affected village implies that these chemicals, at least for the time being, may not be linked to the outbreak. Though statistically not significant, similar rate of difference is observed in terms of storage of grains for six or more months. Therefore, aflatoxin toxicosis may similarly be ruled out unless laboratory tests on the various samples present evidence to the contrary. Special attention, however, needs to be drawn to the dangerous practices of using potentially toxic or noxious substances such as the seeds of Argemone mexicana, melted polyethylene plastic bags, and leaves of wild plant species by women for different purposes including cooking in the affected village. No doubt that such undertaking could in the long-run expose the community to serious health hazards, including liver complications. In light of the foregoing discussion and the over all field survey profile, there is acute need to focus on potentially toxic chemical substances that could possibly be responsible to exposure of the community to the disease in the affected village.

Thus it is deemed necessary to conduct laboratory analysis on various samples collected from case as well as control sites for the presence of toxic substances in order to facilitate rapid intervention approaches. The claim of use of pesticides in stored grains by the respondents in the control village is dismissed as baseless by the woreda agricultural experts who participated in the FGD. According to this group of experts the use of pesticides is not only uncommon but also strictly regulated by the agricultural development agents. The two sites differed significantly with respect to their water sources used for drinking. The higher percent (97%) of the respondents in the case village rely on unprotected or protected well compared only to 3.3% in the control village. This difference is highly significant p=0.000 (OR=840, 95% CI = 50.1, 14095.7). The unprotected or open well is ideal ground not only for accumulating environmental or man made pollutants but also for reproduction and multiplication of pathogenic organisms or even for hydrophilic higher plant species. The toxic constituents of some of these species are undoubtedly get extracted by hydrolysis into the water body where their concentration accumulates in the stationary water. This is possibly due to the formation of aminoalcohol moiety (necine base) and the acid moiety from structural part of PA, of which the alcohol moiety (necine base) is highly water-soluble (17). The high ambient temperature as noted in the affected village combined with sustained disturbance/ agitation of the water by people and other animals easily promotes the process of such extraction. Arial parts of many plants in the immediate vicinity of the water source are often harvested by women and children for washing up containers, especially the widemouthed earthen pots before filling and as a cover afterwards. The crushing of the plant and pressure created in due course of washing the wide-mouthed earthen pots facilitates the release of chemicals including probably toxic ones both to the water in the well and into the potted water. Though the growth of most hydrophilic higher plants in and around drinking water sources does not pose serious health problem to humans or animals, PA containing plants in such surroundings could be the prime suspect in situations like Tsaeda Emba where localized liver disease of unknown origin has been circulating for a number of years. The variation between the two sites with respect to water sources provides an important clue for rigorous physico-chemical analysis and laboratory animal

ACKNOWLEDGMENTS
We express our deep appreciation to the community members of Tsaeda Emba and Jek Emba villages, members of the health office of Tahatay Koraro woreda and officials of the health bureau of Tigray Regional state. We would also like to extend our gratitude to data collectors, FGD participants, key informants and others that were directly/indirectly involved in the investigation. Financial and technical assistance for this investigation was provided by the Federal Ministry of Health, Ethiopian Health and Nutrition Research Institute and the Drug Administration and Control Authority.

feeding trials of the water samples collected from these sites along with samples of plants found growing in or around the unprotected open well. The outcome of these laboratory analyses are presented elsewhere in this edition of the Journal. In light of the slight but statistically insignificant difference between the two sites in terms of the other exposure risk factors considered during this study, the water from the unprotected/protected well seems to be the major culprit for exposure to the disease in the affected village. This, however, also needs to be followed up and substantiated with further longitudinal laboratory study on environmental as well as biological samples to be collected every three or so months.

REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Tsega, E.. Current views on liver diseases in Ethiopia. Ethiopian Medical Journal. 1977; 15:75-82. Boshankova, T. Gillum, R.L. Cirrhosis of the liver in 100 Ethiopians. Ethiopian Medical Journal. 1972; 10:139-44. Lester, FT. Tsega, E. Clinical features of cirrhosis of the liver in 100 Ethiopians. Ethiopian Medical Journal. 1978; 16:13-7. Waldeyes, A. Johnson, O. Mengistu, M. Management of sigmoid volvulus with special reference to primary resection. Ethiopian Medical Journal. 1976; 14:143-50. Kubusta, M.. Fresh tissue examination of liver biopsy specimens in Schistosoma mansoni infection. American Journal of Clinical Pathology. 1965; 44: 283-9. Bergner, P. P. symphytum: Comfrey, coltsfoot, and pyrrolizidine alkaloids. Medical Herbalism,1(1):3-5. Available at: http://medherb.com/articles.htm. Accessed on 11/11/06). WHO on line publication.: www.emro.who.int/Publications/EMHJ/0401/22.htm. Accessed on 15/15/06. Pineiro-Carrero,V.M.and Pineiro,E.O:http://pediatrics.aappublications.org/cgi/content/full/113/4/ S1/1097.Acessed on 11/11/06) Team Members of Faculty of Medicine, AAU and Regional Health Bureau Report on Investigation of an outbreak on an unspecified liver disease in Tahatay Koraro District, Tigray Administrative region, Dec. 2005. Baltazar, JC. The potential of case control method for rapid epidemiological assessment. WHO Statistics Quarterly. 1991; 44 (3):140-4. Robin, S. Galbraita, M. Outbreak of Enterotoxigenic E. coli (ETEC) in Swift Current. Saskatchewan Epidemiology Report. 1995; Vol. 1, No 3. Mattocks, AR. and Jukes, R. New improved tests for toxic pyrrolizidine alkaloids, Journal of Natural Product. 1981; 50: 161-6. Cuniff, P. Official Methods of Analysis of American Official Association of Analytical chemistry (AOAC) International, 16th Ed, Vol. II, AOAC International, 1995; Virginia. WHO Guidelines for drinking-water quality, 3rd Ed., Vol. II, 2004 Geneva. Chorus, I and Bartam, J. Toxic cyanobaceria in water, Spon press, 1999, London. Loomis, T.A. Essential of toxicology, 3rd. edition, Lea Fabiger, 1978, pp190-8. Mattocks, A.R.. Hydrolysis and hepatotoxicity of retronecine diesters. Toxicol. Lett., 1982, 14: 111-6.