Poisoning of Kurdish refugees in Turkey


Kurdish Scientific and Medical Association, London WC1


Poison Unit, New Cross Hospital, London


Department of Chemical Pathology, Old Medical School

University of Leeds, leeds LS2 9JT, UK

THE LANCET, 1990: VOL 335, 287/8

SIR- Poisoning by an unknown agent was reported in June, 1989, to have affected some 2070 Kurdish refugees (667 children, 740 women, 663 men) living in Mardin camp in Turkey (total population 15 157)1. The refugees had fled from Iraq in 1988 to escape government forces harrying them with chemical and conventional weapons. The symptoms were diarrhoea, abdominal cramps, vomiting, disturbance of speech, disorientation, inability to walk on a straight line, general weakness, and temporary paralysis of the limbs; recovery was slow. The symptorns suggested some neurotoxic agent(s). Samples of blood and bread (thought to have been the vehicle for the poison) were brought back to the UK by J. F. and by Mr Gwynne Roberts, a television journalist.

Blood was taken from 20 very sick men, women, and children (aged 2?50) five days after the occurrence of symptoms. Only 8 samples could be transported to the UK. Blood was put in sterile heparinised plastic tubes and kept at ambient temperatures for five days before analysis. Qualitative analysis for neurotoxic heavy metals was done in the clinical biochemistry departments of the University of Southampton, St Luke's Hospital, Guildford, and in the poison unit, Guy's Hospital Medical School. Thallium, mercury, lead, copper, and barium were sought by inductively coupled plasma source mass spectrometry. One laboratory reported lead, mercury, and barium in one sample of bread and in two blood samples; the other two laboratories could riot confirm those findings in the same, or other, samples. Screening for trichothecene mycotoxins in bread was done after extraction and elution on charcoal?aluminium columns2. No mycotaxins were detected. Screening for nitrogen/phosphorus compounds was done by gas chromatography and gas chromatography/mass spectrometry.

Cholinesterase activity was assessed in two blood samples by measuring the changes in absorbence at 405 nm (EPOS 5060 analyser) resulting when the chromogen 5,5'?dithiobis(4-nitrobenzoic acid) reacts with the thiocholine iodide produced by the action of the enzyme on a highly diluted sample of acetylthiocholine iodide. The assay was calibrated with acetylcholinesterase of known activity The two samples evinced severe inhibition, the lower Haut of normal for cholinesterm being about 9.0 kU/l:

To ensure that high ambient temperatures had not reduced the cholinesterase activity, a standard solution of pure enzyme was kept at 25oC for five days. The enzyme activity in this sample did not differ from that of a freshly thawed standard. To exclude genetic or nutritional causes for the low enzyme activity, a cholinesterase reactivator, prallidoxime (4 pg per 500 ml blood), was introduced to reactivate the cholinesterase in the two samples and in the controls. Prallidoxime will also remove organophosphates attached to the activity site of the cholinesterase causing an increase in enzyme activity. A three?fold increase in activity was recorded in the samples whereas the increase was only 20% for the controls.

Gas chromatography of the test blood samples revealed only one significant peak, indicating a nitrogen/phosphorus compound. This was identified as 2-hydroxyethylbenzthitzol, a product of ethylene oxide (a sterilising agent) and benzthiazol (a vulcanising agent), and was almost certainly present in the syringes used to collect the samples. In control samples a large concentration of 2-hydmxyethylbenzthiazol (200 mg/ml) had no detectable effect on cholinesterase activity either immediately or after five days' incubation at room temperature.

A sample of bread was extracted into diethyl ether and the solvent was evaporated to dryness. Some of the residue was then added to two reference blood samples. There was only a very small and not significant decrease in cholinesterase activity.

We would have liked to have had more blood samples but the 8 analysed were smuggled out at great risk, and the Turkish authorities would not allow independent medical investigators into the camps.

The symptoms reported were consistent with poisoning by a neurotcoxic agent, but there were very many possibilities so the three laboratories were asked to look for particular agents. The finding of heavy metals in some samples by one laboratory could not be repeated by the other two, which suggested that metals, if present, were not at concentrations sufficient to cause the symptoms. Valuable material was used up in the screening process, and it was only as a last resort that the two remaining specimens were tested for cholinesterase inhibition. The findings point to a potent nerve agent (organophosphorus) as the cause of the poisoning.

Commercially available organophosphorus pesticides are an improbable source of the poisoning because of their low toxicity. Their foul smell and taste would make it difficult for anyone to consume sufficient to cause the symptoms reported. Organophosphorus pesticide metabolites are quite easy to detect but none were found by the UK National Poison Unit. Nerve gas cannot be ruled out. They would cause some of the symptoms reported, but some equally toxic organophosphate may have been responsible. Organosphosphorus chemical warfare agents inhibit acetylcholinesterase and do so rapidly. The rate of recovery is determined by the degree of poisoning and by what caused it, and enzyme activity can remain depressed for weeks.

Inhibition of acetylcholinesterase would explain the symptoms reported by most of the patients in Mardin, and the rate of recovery. The fact that no traces of an organophosphate agent were found might be explained by the time delay between poisoning and blood collection. Bread had been singled out as the most likely source of the poisoning but we can find no evidence to support this claim. The bread sample we received may not have been representative or a toxic organophosphate originally present in the bread may have been inactivated. No detailed Investigation in the camp was done and the source of the cholinesterase inhibiting agent remains unknown.

These refugees had fled from chemical attacks on their homes in northern Iraq in August, 1988. They now claim to have been the victims of another mass?poisoning attempt. The evidence available does suggest something very sinister. It is unlikely that we are talking about a common commercially available chemical, so the chance of accidental poisoning is remote. Most of the victims have recovered, and the UN High Commission for Refugees now has access to Kurdish refugee camps in Turkey. If this poisoning was, as we strongly suspect, deliberate, it has serious implications for the international community.

We thank Dr J. Henry and Dr B. Widdop (National Poison Unit, Guy's Hospital); Dr A. Walker and Dr A. Taylor (St Luke's Hospital, Guildford), and Dr T. Delves and Mr C. Fellows (Southarripton University) for help and advice.


1. Independent June 2,1989.

2. Thomas R, Ronier J. The use of small charcoal/alumina clean?up columns in the detection of tricothecene mycotxins in foods and feeds. J Assoc Off Analyt Chem, 1986; 69, 699-703.