Asfaw Debella, Dawit Abebe, Frew Tekabe, Ambaye Degefa, Assefa Desta, Alemtsehaye Tefera, Aberra Geyid, Kissie

Mudie, Ashenife Tadele, Derebew Getahun, Teklele Biza, Tesfaye Kebede. 2011. Ethiop Med J, Supplement 1

PHYSICO-CHEMICAL INVESTIGATION OF CONSUMABLES AND ENVIRONMENTAL SAMPLES TO DETERMINE THE CAUSATIVE AGENT OF LIVER DISEASE OUTBREAK IN TAHITAY KORARO WOREDA, TIGRAY
Asfaw Debella1, Dawit Abebe1, Frew Tekabe1, Ambaye Degefa1, Assefa Desta1, Alemtsehaye Tefera1, Aberra Geyid1, Kissie Mudie1, Ashenife Tadele1, Derebew Getahun1, Teklele Biza1, Tesfaye Kebede1.

ABSTRACT
Background: The occurrence of liver disease out break of unknown origin and of a significant morbidity and mortality with a major symptom of abdominal dropsy was reported among the inhabitants of Tsaeda Emba village, Tahtay Koraro woreda, Tigray during mid December, 2005. Objectives: The aim of this study was to assess if chemical intoxicants were the responsible agents for the outbreak in the affected locality. Methods: The experimental approach towards the identification of the possible causative agent from the consumables and environmental samples collected from the affected village were determination of physico-chemical quality parameters and chemical analysis using instrumental and chromatographic techniques. Results: The intoxicant was found to stem from the consumption of the water from the unprotected well in which pyrrolizidine containing plant, Ageratum sp, abundantly thrives. Conclusion: The water source was intimately linked to the outbreak of disease and, therefore, immediate intervention and preventive measures are called for. Key words: Unspecified liver disease, intoxicants, physico-chemical analysis, unprotected pond plants affecting livestock, wildlife and humans. Intoxication commonly occur through the use of these plants for medicinal purpose, or as a contaminants of agricultural crops, and by consumption of honey, egg or milk which are produced by animals that forage on pyrrolizidine-containing plants (4-6). Incidence of veno-occlusive disease is frequently observed in countries with poor agricultural systems as a result of grain contamination by plants containing pyrrolizidine alkaloids (7-9). Ingestion of foods contaminated with certain species of fungus toxins (aflatoxins) even in minute amounts over a long period of time can also cause acute or chronic liver disease. Aflatoxicosis outbreak was reported from India and Kenya as a result of consumption of contaminated grain (10,11). In addition to the above mentioned organic compounds, acute or chronic nature of toxic hepatitis can also occur as a result of excessive contamination of the soil and ground water by inorganic compounds such as arsenic, selenium, beryllium, cadmium and copper (12,13). Consump-

INTRODUCTION

A number of factors contributes to the liver disease, some of the major ones that induce toxic hepatitis include ingestion of consumables such as grains, edible oils, etc. that may have been adulterated or contaminated by plants with toxic metabolites. One of such plants is Argemone mexicana which elaborates the toxic alkaloids, sanguinarine and dehydrosanguinarine causing hydropic change and degenerated nuclei of the hepatocytes (1). An outbreak of epidemic dropsy has been reported in India and South Africa due to the contamination of wheat flour with seeds of A. mexicana (2, 3). Plants containing pyrrolizidine alkaloids are the other most common causes of hepatotoxicity. Prolonged exposure to these compounds leads to veno-occlusive disease as a result of obstruction of the small veins that bring blood from the liver back to the heart. They are poisonous
1

Ethiopian Health and Nutrition Research Institute, Addis Ababa, Ethiopia, P. O. Box 1242, Addis Ababa, Ethiopia. .

tion of water containing high concentration of cyanobacteria or ingestion of the toxin during swimming can also cause a range of hepatotoxic symptoms and necrosis (14). An outbreak of human poisoning attributed to toxic cyanobacteria has been reported from Australia, United Kingdom and Brazil (15). An outbreak of liver disease of unknown origin in Tahtay Koraro that could have possibly arisen due to one of the above causes was reported in mid December, 2005 (16). On the basis of the report, EHNRI was instructed by the Federal Ministry of Health to investigate the situation surrounding the affected locality. In response to this call, multi-disciplinary task force was set up to investigate the cause of the out break. The task force identified a number of activities that need to be undertaken to rule out or rule in whether the fatal liver disease is of toxic or of infectious origin. Following the review of the initial report, immediate laboratory analysis was conducted on the consumable samples collected from the affected area and submitted along with the report. The fact that the submitted samples were mostly inadequate and of poor quality, the task force decided to undertake its own field trip to the affected area to collect epidemiological information and samples of biological and environmental nature. In view of the above mentioned predisposing factors to heptotoxicty and the correlation of the major symptoms that appear to be largely linked to noninfectious intoxicants and the localized nature of the problem, it was felt that there is a need to give more emphasis in detecting the presence or absence of organic and inorganic intoxicants during the laboratory investigation. Hence, the objective of this study was to carry out detailed physico-chemical analysis on the collected consumable and environmental samples in order to facilitate the identification of the causative agent (s) that might have been responsible for the outbreak in the affected village.

Sample types: Water, cereal samples of millet, sorghum, maize and Teff,, and Tela samples from case and control villages; samples of cooked meal, mainly Injera from the case village; samples of cooking oil from the market; plant samples that are found floating/submerged in the unprotected open water well; water samples from un protected spring in the control village as well as the unprotected pond and well fitted with a hand pump from the case village; wild edible plants and plants used for different purposes (plant leaves kept as a mat under Injera, or used as wrappers during the germination process of barley, bikil). Laboratory tests for physico-chemical quality parameters and chemical analysis of samples for toxic metabolites: Samples of cereal flours, water, cooking oil and cooked meal were taken and extracted with appropriate solvents before being subjected to various test reagents and TLC analysis (17-19). The samples of water and cooking oil were further analyzed for physico-chemical quality parameters and toxic inorganic elements according to the method outlined in APHA, AWWA and WEF (19, 20).The water collected from the unprotected well of Mai-habi-tselim was also tested for cyanogenic bacteria or the toxic products thereof according to standard methods (21,22).

RESULTS

Chemical analysis of consumable samples: Tests for organic intoxicants, viz., pyrrolizidine alakaloids, tannic acid, aflatoxins and organo-chlorine pesticides were negative in samples of cooked foods from the case area. The same result was recorded for flour samples of millet, sorghum, maize and Teff collected from the case and control villages. Although trace PA is detected form the water samples of unprotected pond from the case site, analysis of the water samples form control areas, tela and cooking oil samples from the case area did not show the presence of toxic plant metabolites ( Table 1).

MATERIALS AND METHODS

Sampling technique: Sample collection technique is generally based on case-control method which is described elsewhere in this edition of the Journal.

Table 1 Tests for cyanogenic bacteria and chemical analysis on water, tela and cooking oil samples collected from affected and non-affected villages No. Sample Test for pyrrolizidine alkaloid Test for tannic acid Test Argemone mexicana alkaloids Test for cyanogenic bacteria Remarks

1 1.1 1.2 2 2.1 2.2 3 3.1 3.2

4 4.1 4.2 5

Water (Tseada Amba) Well fitted with a hand pump Unprotected pond Water (Jek Amba) Well fitted with a hand pump Unprotected spring water Cooking oil Shops where farmers commonly buy Shops where city dwellers commonly buy Home brew Tela From Tseada Amba village From Jek Amba village Argemone mexicana seed oil alkaloids (control)

Negative Trace

Negative Negative

Negative Negative

Negative Negative

Negative Negative Negative Negative

Negative Negative Negative Negative

Negative Negative Negative

Negative Negative

Not applicable Negative Not applicable

Negative Negative Negative

Negative Negative Negative

Negative Negative Positive

Not applicable Not applicable Marker for Dihydrosanguinarine & sanguinarine alkaloids

The test results of water quality parameters indicated that the water samples from all the four sources are chemically potable according to the guideline values of the WHO (Table 2). There also appears to be no significant variation among the water samples for all the four sources with respect of most of the chemical quality parameters. However, though the presence of cynobacteria could not be established, substrates such as nitrate, nitrite, and the limiting nutrient, phosphate, that favor the growth of cynobacteria were detected 2 to 5 times higher amounts in the water from the unprotected

well of the affected village than in the samples collected from the other three sources (Table 2). Results of inorganic chemical analysis of the water samples from the affected and non-affected villages also revealed a relatively increased (10 times higher than the threshold level of WHO) level of lead in the water from the wells fitted with hand pump that are erected in both villages and from the unprotected well of Mai-habi-tselim in the affected village (Table 3). The physico-chemical characteristics of the cooking oil complies with codex standards of edible oil except for the clarity and very high acid value that may cause gastric irritation (Table 4).

Table 2 Tests on quality parameters of drinking water samples collected from affected and non-affected villages Water from Tseada Amba No Physico-chemical parameters Well fitted with hand pump 6.86 140.0 136.0 354.6 36.87 10.70 0.03 Nil Nil 170.8 12.0 0.03 28.8 0.01 Trace Unprotected well 6.66 146.4 116.0 318.3 30.46 9.73 Nil Nil 0.03 178.61 12.0 0.07 31.95 0.09 Trace Water from Jek Amba Well fitted with hand pump 7.37 146.4 136.0 365.0 32.0 13.6 Nil Nil Nil 178.61 12.0 0.01 43.1 0.04 Trace Unprotected spring 7.60 146.4 84.0 264.0 19.0 8.8 0.03 Nil Nil 178.61 7.2 Nil 18.53 0.015 Trace

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

pH Value Total Alkalinity, mg/L as CaCO3 Total Hardness, mg/L as CaCO3 Electrical Conductivity at 250C, S/cm Calcium (Ca++), mg/L Magnesium (Mg ) Iron (Fe), total Manganese, (Mn ) Ammonium (NH4+) Bicarbonate (HCO3 ), mg/L Chloride (Cl) Nitrite (NO2 ) Nitrate (NO3-) Phosphate (PO4 ) Sulphate (SO4-2)
-3 +2 ++

Chemical analysis of the environmental samples: No accumulation of toxic secondary metabolites were noted as a result of tests conducted on plant samples that were used as ingredients in the preparation of Tela, or as items for household utility and those used as source of wild edible fruits (Table 5).

However, the plant identified as Ageratum sp (local name unknown) that grows submerged or floating in the unprotected well of Mai-habi-tselim in the affected village was found to be strongly positive for the presence of pyrrolizidine alkaloids which was confirmed by positive control (Table 5).

Table 3. Inorganic chemical analysis on water samples collected from affected and non-affected villages Well fitted with hand pump (Tseada Amba) < 0.1 Well fitted with hand pump (Jek Amba) < 0.1 Un protected spring water (Jek Amba) < 0.1 Guide line (GL) value for water2004 (mg/l) NO GL b/c Be is unlikely to occur in drinking water 2 No GL 0.05 (provisional) 0.01 0.003 0.7 0.02 (provisional) 0.01 g/ml 0.01 g/ml

No.

Water quality parameter Beryllium (Be)

Un protected well (Tseada Amba) < 0.1

2 3 4 5 6 7 8 9 10

Copper (Cu) Cobalt (Co) Chromium (Cr) Lead (Pb) Cadmium (Cd) Barium (Ba) Nickel (Ni) Arsenic (As) g/ml Selenium (Se) g/ml

< 0.1 < 0.1 < 0.1 0.1 < 0.1 < 0.1 < 0.1 0.011 0.29

< 0.1 < 0.1 < 0.1 0.1 < 0.1 < 0.1 < 0.1 0.013 Not detected

< 0.1 < 0.1 < 0.1 0.1 < 0.1 < 0.1 < 0.1

< 0.1 < 0.1 < 0.1 <0.1 < 0.1 < 0.1 < 0.1 1.11

sample size was little to do the test

Not detected

Table 4 Physico-chemical characteristics of cooking oil purchased from a market frequented by residents from the affected village No 1 Parameters Description Shops where farmers commonly buy Yellowish- brown, turbid liquid with a significant amount of settleable matter. 0.922 1.468 107.5 192.0 2.76 7.8 Except the clarity and the very high acid value, the other physicochemical characteristics comply with the Codex standard for edible oils. Shops where city dwellers commonly buy Golden- yellow, turbid liquid with a significant amount of settleable matter 0.920 1.469 108.05 191.7 2.65 8.9 Except the clarity and the very high acid value, the other physicochemical characteristics comply with the Codex standard for edible oils.

2 3 4 5 6 7 8

Specific gravity at 200 C Refractive index at 400 C Iodine value Saponification value Acid value Peroxide value in mEq/kg Organoleptic characteristics

Table 5 Chemical analysis on plant samples collected from the affected village No. 1 1.1 1.2 1.3 2 Sample Plant samples that are floating in the water or nearby open well Buddleja filisita [Mehemar ketele (Tg)] Laggera alata [Dega tsela (Tg)] Ageratum sp. [local name is not known] Plants that are kept under Injera (in mesobe or while serving food) to prevent it from drying and in covering during the germination process of barley (bikil) Diospyros mespiliformis [Ayu ketele (Tg),] Rhus natalis [Tatelo (Tg)] Dombeya sp. [Kesela ahewie (Tg)] Croton machrostachyus [Medfie ketle (Tg), Bisana (Am)] Wild edible plants Ripe and unripe seeds of Bangueria edulis [Guramaile (Tg)] Ripe and unripe seeds of Diospyros mespiliformis [Ayu fre (Tg)] Positive Controls Senecio gigas leaves, flowers & stem [Yeshikoko gomen] Tannic acid Test for pyrrolizidine alkaloid Test for tannic acid Remarks

Negative Negative Positive

Negative Negative Negative

Possible intoxicant

2.1 2.2 2.3 2.4 3 3.1 3.2 4 4.1 4.2

Negative Negative Negative Negative

Negative Negative Negative Negative

Negative Negative

Negative Negative

Positive Negative

Negative Positive

Marker for pyrrolizidine alkaloid Standard chemical

DISCUSSION
The absence of pyrrolizidine alkaloids, tannic acid, aflatoxins and organo-chlorine pesticides in the cereals and cooked meals showed that the consumables were not contaminated with the toxic metabolites of plants or fungi. The absence of aflatoxin in the grain samples collected from 20 households was in line with the expectations since these were not obtained from underground storage facilities nor were they stored for a long time. Furthermore, the use of pesticides and commercial fertilizers at the time of sample collection were also uncommon in the affected area, which minimizes the possibility of toxicity as a result of contamination of the grains by chemical or pesticide residues.

No toxic substances were detected in the consumable samples despite the admission of the women FGD participants of the use of poly ethylene plastic sheets (festal) that generate dioxin during combustion, seeds of castor plant and of Argemone mexicana which produce toxic ricin and toxic alkaloids, respectively, during injera baking or preparation of tela. The small sample size and/or the low level of sensitivity of the laboratory procedure may have contributed to the non-detection of these intoxicants. The use of seeds of A. mexicana to increase the inebriating strength of Tela in the affected village is a very dangerous practice that could lead to serious health hazards including abdominal dropsy as reported in other countries. Sanguinarine and dehydrosanguinarine, the toxic constituents of the seeds of A. mexicana reported to cause an outbreak of epi-

demic dropsy in India and South Africa (1, 3).Ricin, the toxic substance from the residue of the seeds of Ricinus communis, commonly known as castor bean comprises 2.8 to 3% of the total weight of the seeds could prove lethal to any person (23). The presence of ammonia, nitrite, nitrate and phosphate suggests the existence of favorable environment condition for the potential growth of algae or cyanobacteria in the open well. They are also good indicators of low sanitary quality of the water source. Besides, the greenish and turbid nature of the water and the abundant presence of animal excreta around the well that could be washed into the water may also favor algal growth. This observation is in line with literature reports that point out problems associated with cyanobacteria are also likely to increase in areas with lack of concomitant sewage treatment and with agricultural practices causing nutrient losses to water bodies through erosion (24). The negative result for cyanobacteria in the sample of water from the unprotected well may be attributed to the collection period or season. It is a common knowledge that cyanobacteria or algal bloom requires optimal conditions of temperature, light and nutrient status (24). Drinking of water containing high concentration of toxic cyanobacteria or their toxins if causes serious hazard to human health (14) and outbreaks of human poisoning attributed to toxic cyanobacteria have also been reported in various countries (15, 25). The presence of Ageratum sp in the water catchments site of the affected village was in agreement with the localized outbreak of veno-occlusive liver disease since many species belonging to the genus elaborate toxic pyrrolizidine alkaloids (PA) (26, 27). These secondary plant metabolites are reported to be hepato -toxic affecting livestock, wildlife and humans. The presence of this procumbent herb found in and around the unprotected well means that its pyrrolizidine alkaloids could freely be extracted into the water body. Trace PA was detected from the suspected water sample (un protected pond), in addition to this the metabolite of pyrrolizidine alkaloid (pyrrole) was also detected during the chemical analysis of the extracts of the liver specimen of sacrificed mice following feeding trial (giving as a drink) of the water on laboratory animals. This indirectly confirms that the suspected water that possibly contained the toxic substance, PA could be a cause for intoxication

thereby leading to the outbreak. Furthermore, the aerial part of the plant may also be collected to be used for washing the containers before filling and/or as a cover after filling which is a common practice in most rural areas, a practice which inadvertently facilitates the release of PA from the injured plant. Besides the above possible explanation, the community in the affected village might also use other PA containing plants for different ailments including for the reported unidentified disease. In fact, the participants of the FGD at the community level in the affected village mentioned the use of powdered roots of Lemiene bahata (Tigrigna) and stem bark of Weyebo (Tigrigna) dissolved in milk for the treatment of malaria, gastritis and other ailments. The survey team could not get a voucher specimen of these plants for taxonomic identification and chemical analysis, since these plants were reported to be found very far from the affected area. Therefore, in the present situation it is very difficult to consider some possible health hazards associated with these traditionally used medicinal plants unless the information obtained is complete and justified by an experiment. On the basis of the present laboratory investigation, the intoxicant responsible to the stated health problem is most likely pyrrolizidine alkaloid and to a lesser extent perhaps toxins of cyanogenic bacteria. The water from the unprotected well of Mai-habitselim is considered the major source for the cause of the disease because of the growth in it of the plant Ageratum sp that elaborates the hepatotoxic pyrrolizidine alkaloid. Besides the greenish and turbid nature of the water from the same well coupled with the presence of nitrate, nitrite and the limiting nutrient phosphate in substantial amount is indicative of a favorable environment for the multiplication of cyanogenic bacteria (algae) whose toxins are considered as one of the predisposing factors for venoocclusive liver disease. Furthermore, although the water from the well fitted with a hand pump is grossly clear and clean to the naked eye, the risk of contamination of this water by the unprotected pond is very high because of leachable toxic substances, as both wells are located down hill and are in close proximity of about 3 meters. Conclusion: The water source was indirectly linked to the outbreak of disease due to consumption of the water from the unprotected pond of Mai-habi-tselim in which pyrrolizidine alkaloid containing plant Ageratum sp. abundantly present. In order to control the

outbreak and avoid further exposure to the intoxicant the unprotected well of Mai-habi-tselim and the well fitted with the hand pump in the affected village should be barred immediately and alternate safe and potable water source should be made available to the residents. As a preventive intervention strategy, extensive health education should be given to the community with regard to dangerous practices such as using potentially toxic plants or other noxious substances so as to minimize the occurrence of similar episodes. There is also a need to put in place a disease surveillance system focusing on non-infectious diseases. Documentation of information with regard to environmentally hazardous substances, plants that are either used medicinally or considered to cause health havoc to human or livestock is of a paramount importance to facilitate the identification of noxious and toxic substances and to monitor as well as to increase

awareness among health professionals and the general public.

ACKNOWLEDGMENTS
We are very grateful to Federal Ministry of Health, Ethiopian Health and Nutrition research Institute and Drug Administration and Control Authority for financial and technical support for the investigation. We express our deep appreciation to the community members of Tsada Emba and Jek Emba villages, members of the health office of Tahtay Koraro woreda and officials of the health bureau of Tigray Regional state. Our acknowledgment also goes to data collectors, FGD participants, key informants and others that were directly/indirectly involved in the investigation

REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Sarkar SN. Isolation from argemone oil of dihydrosanguinarine and sanguinarine: toxicity of sanguinarine. Nature. 1948; 162:265. Ramasastri BV, Babu S. A study on the toxicity of Argemone oil in experimental animals. Indian Journal of Medical Research. 1975; 63:1353-7. Singh GB, Khanna SK. Argemone mexicana strikes again. Science Reproduce. 1983; 20:108-110. Weston C. Veno-occlusive disease of the liver secondary to ingestion of comfrey, British Medical Journal. 1987; 295: 183. Datta DV. Herbal medicines and veno-occlusive disease in India, Postgraduate Medical Journal. 1978; 54 (634): 511-5. Roulet M. Hepatic veno-occlusive disease in newborn infant of a woman drinking herbal tea, Journal of Pediatrics. 1988; 112(3): 433-6. Datta, DV., Khuroo, MS., Mattocks, AR., Aikat, BK., and Chhuttani, PN. Veno-occlusive disease of liver due to Heliotropium plant used as medicinal herb (report of six cases with review of literature). Journal of Association of Physicians of India. 1978a; 26(5): 383-93. Grases, PJ. and Beker, SG. Veno-occlusive disease of the liver. A case from Venezuela. Am. J. Med. 1972; 53: 511-6. Hooper, PT. Pyrrolizidine alkaloid poisoning-pathology with particular reference to differences in animal and plant species. In: Keeler, R.F., Van Kampen, KR. and James, LF., ed. Effects of poisonous plants on livestock, New York, Academic Press, 1978; pp. 161-76. Wyllie, TD. and Morchause, LG. Mycotoxin Fungi, Mycotoxins, Mycotoxicoses-An Encyclopedic Handbook.Vols.1,2, and 3.Marcel Dekker, Inc. 1978; New York. Heathcote, JG. and Hibbert, JR.. Aflatoxins: Chemical and biological aspect. Elsevier, New York, 1978; pp.173-86. Guha Mazumder DN. Arsenic levels in drinking water and the prevalence of skin lesions in West Bengal, India. International Journal of Epidemiology. 1998; 27: 8717. Rouiller, C. Experimental toxic enjury of the liver. In: Rouiller, C (ed): The Liver, Vol. 2, Academic Press, Inc., New York, 1964; pp 335-476.

14. Falconer, I.R. Measurement of toxins from blue-green algae in water and foodstuffs. In: I.R. Falconer [ed.] Algal Toxins in Seafood and Drinking Water. Academic Press, London, 1993; 165-75. 15. Baker, PD. and Humpage, AR. Toxicity associated with commonly occurring cyanobacteria in surface waters of the Murray-Darling basin, Australia. Australian Journal of Marine Freshwater Research. 1994; 45: 773-86. 16. Faculty of Medicine, Addis Ababa University, Investigation of an outbreak of unspecified Liver Disease in Tahtay Keraro Woreda, Tigray Administrative Region, December 2005. 17. Mattocks, AR. Detection of pyrrolizidine alkaloid on TLC, Journal of Chromatography. 1967B; 27: 505-8. 18. Mattocks, AR. and Jukes, R. New improved tests for toxic pyrrolizidine alkaloids, Journal of Natural Product. 1981; 50: 161-6. 19. Cuniff, P. Official Methods of Analysis of American Official Association of Analytical chemistry (AOAC) International, 16th Ed, Vol. II, AOAC International, 1995; Virginia. 20. American Public Health Association (APHA), American Water Works Association (AWWA) and Water Environment Federation (WEF) Standard method for examination of water and waste water, 18th. Ed., Victor Graphics, Inc., 1992; Washington, DC. 21. WHO Guidelines for drinking-water quality, 3rd Ed., Vol. II, 2004 Geneva. 22. Chorus, I and Bartam, J. Toxic cyanobaceria in water, Spon press, 1999, London. 23. Duke, JA. Ricinus communis (Euphorbiaceae): In handbook of Medicinal Herbs, CRC Press Inc. 1985C, 408. 24. Siegrist, H. and Boller, M. Effects of the phosphate ban on sewage treatment. EAWAG News. 1997; 42 E: 911. 25. Jackim, E. and Gentile, J. Toxins of a blue-green alga, similarity to saxitoxin. Science. 1968; 162: 915-916. 26. Roder, E, Medicinal plants in Europe containing pyrrolizidine alkaloids, Pharmazie. 1995; 50: 83-98. 27. Zhao XL, Chan MY, and Ogle CW, The identification of pyrrolizidine alkaloid-containing plants, American Journal of Chinese Medicine. 1989; 17 (1-2): 71-8. 28. Culvenor CC. Hepato- and pneumotoxicity of pyrrolizidine alkaloids and derivatives in relation to molecular structure, Chemical and Biological Interactions. 1976; 12(3-4), 299-324.