Bungarus caeruleus (Common Krait) bites are a continuing medical problem in Sri Lanka (^{Kularatne, 2002}; ^{Ariaratnam et al, 2008}). These bites rapidly lead to a classical occulo-facio-bulbar and respiratory muscle paralysis that requires mechanical ventilation in about half of all victims (^{Kularatne, 2002}).

Physicians managing Bungarus caeruleus bites frequently encounter patients in states of deep coma mimicking brain death. A prospective study of 210 cases of Bungarus caeruleus bites in the North Central Province of Sri Lanka reported a high incidence of deep coma (^{Kularatne, 2002}). This series reported a state of ‘total unresponsiveness to deep pain or loud voices associated with fixed dilated pupils and absent brain stem and spinal reflexes' in 35 (17%) patients. All of the survivors in this group (22 patients) made a complete neurological recovery after varying periods of mechanical ventilation. Personal communication with many practicing physicians in the Island revealed similar transient and completely reversible deep coma in a significant number of victims of common Bungarus caeruleus in Sri Lanka. This phenomenon may be due to hypoxia secondary to respiratory failure or a direct effect of venom on the cortical and/or brainstem neurons. We assumed that EEG and brainstem auditory and visual evoked potentials will be abnormal if the venom acted on the cortical and brainstem neurones.

Another intriguing phenomenon reported following Bungarus caeruleus bite in Sri Lanka is hypokalaemia (^{Sellahewa, 1997}; ^{Kularatne, 2002}). Kularatne reported a high incidence (71%) of hypokalaemia (<3.5mmol/l) in their series (^{Kularatne, 2002}). Author postulated beta adrenergic stimulation and intra-cellular shift of potassium as the potential cause of this phenomenon. A third possibility is loss of potassium in the renal tubules. We postulated that measurement of urinary potassium excretion would be a reliable surrogate marker of renal conservation of potassium. If there is intracellular shifting of potassium, the renal tubules would attempt to conserve potassium and thus, potassium excretion of would be minimal. A third possibility is the loss of potassium in the renal tubules.

There are no published reports on human studies conducted to describe the pathophysiology of these phenomena. The objective of the current report is to postulate two putative mechanisms of deep coma and hypokalaemia following Bungarus caeruleus bites in Sri Lanka.

Two patients who were transferred in November 2008 to the study hospital following proven bites by Bungarus caeruleus were prospectively followed up until death or discharge. Regular findings of clinical observations of cardiovascular and nervous systems were documented. Daily blood samples were analysed for serum potassium and sodium while the daily urine was collected for the analysis of potassium excretion. Digital electroencephalograms (EEG) and electrocardiograms (ECG) were performed on a daily basis. In patient 1, we performed peripheral nerve conduction studies (NCS), repetitive nerve stimulation studies (RNS), visual (VEP) and brain stem auditory evoked potentials (BAEP) to assess the peripheral and central nerve pathways. Clinical management included replacement of potassium at a rate of 1-1.5mmol/kg/day, replacement of fluids and nutrients.

Both patients were presented to the local hospital with the killed specimen of Bungarus caeruleus. Both patients were treated with polyvalent antivenom (20 vials of Lyophilised, Enzyme refined, Equine Snake Venom Antiserum; (Vins Bioproducts Limited^® ) raised against Indian Cobra, Common Krait, Russell's Viper and Saw Scaled Viper) at the local hospital, before being transferred to the study hospital for further care.

These two cases demonstrate clearly that severely envenomed patients with Bungarus caeruleus have depressed cortical activity and brain stem functions. In addition to deep coma, both patients had identical EEG activity of absent alpha and delta waves and dominant theta waves after the bite. The survivor had restored alpha activity on the third day while in the other, alpha activity did not return. Both patients demonstrated hypokalaemia with reduced excretion of urinary potassium. Replacement of potassium was unable to improve serum potassium or urinary potassium excretion. The cause of death in the Patient 1 was a combination of disseminated intravascular coagulopathy and adult respiratory distress syndrome (ARDS). The latter complication has been reported previously by Kularatne (^{Kularatne, 2002}) in 5 patients who died while being in a deep coma. Necrosis of diaphragmatic muscles following Bungarus caeruleus has not been reported previously and may have contributed to the respiratory failure.

Venom of Bungarus caeruleus contains a mixture of alpha- and beta-bungarotoxin and a caerulotoxin. Beta-bungarotoxin acts on the presynaptic terminal while alpha-bungarotoxin and caerulotoxin act on the post synaptic membrane (^{Abe et al, 1977}; ^{Bon and Changeux, 1977}). Beta-bungarotoxin of Bungarus multicinctus causes rapid release of acetyl choline (ACh) from the pre-synaptic membrane, but rapidly causes inhibition of choline re-uptake into the pre-synaptic membrane thus depleting choline available for recycling of ACh. Therefore, ACh levels are decreased leading to an inhibition of neurotransmission (^{Sen et al, 1976}), which causes peripheral neuromuscular paralysis. In addition to its effects on the peripheral nervous system, beta bungarotoxin has been shown to possess the ability to act on the central cholinergic systems of the cortex, cerebellum and hippocampus in rats (^{Gulya et al, 1984}). Inhibition of central cholinergic neurotransmission may theoretically cause deep coma with representative EEG abnormalities. In the rat, the alpha waves are generated from the thalamocortical pathways, predominantly in the sensory motor cortex (^{Semba and Komisaruk, 1984}). In the Rhesus, monkey, alpha-bungarotoxin binds to the cortex (Han et al, 2003). Both patients presented in this report had absent alpha activity, which gives indirect evidence of inhibition of cortical neurotransmission. Theta activity is generated from the hippocampal area in animals. Alpha-bungarotoxin binds extensively throughout the hippocampi of both the rat and the Rhesus monkey (^{Freedman et al, 1993}; Han et al, 2003). In the rat, binding is particularly heavy with GABAergic interneurones in the dentate gyrus and the Ammon of the hippocampus (^{Freedman et al, 1993}). Theta activity is believed to be generated in the hippocampus in both animals and man (^{Miura et al, 1985}). This is further strengthened by the fact that addition of scopolamine, a acetylcholine inhibitor, inhibits the generation of theta activity in the rat hippocampus (^{Brazhnik et al, 1993}). Excess theta activity observed in both patients perhaps indicates that bungarotoxin does not bind, or does so less avidly, in the humans. In patients with other forms of coma, such as cerebral malaria, delta activity, a waveform believed to be generated from the thalamus in association of the reticular activating system (RAS), is the predominant wave form observed in the EEG (^{Thumasupapong et al, 1995}). Neither of the patients showed delta activity indicating that a mechanism different to that observed in metabolic coma is in operation in patients with Bungarus caeruleus bite-induced coma.

Further, beta-bungarotoxin from Bungarus multicinctus has been shown to cause release of the inhibitory neurotransmitter, GABA, in experimental models (^{Wernicke et al, 1975}), which may also contribute to deep coma. Patient 2 and all the severely envenomed patients reported by Kularatne (^{Kularatne, 2002}) had anterograde memory loss, a phenomenon that further supports reversible inhibition of these structures.

There may be dose-dependent structural damage to the synapse membranes, which may explain the lack of recovery in one of the patients presented (^{Ueno and Rosenberg, 1996}; ^{Herkert et al, 2001}). Preferential binding of bungarotoxin to the brain stem structures is the most likely explanation of absent brain stem reflexes and VEP and grossly delayed BAEP. Involvement of brain stem should be a key contributory factor in the rapid the development of respiratory failure.

Ramachandran reported EEG abnormalities of a heterogeneous group of patients with snake bites in Sri Lanka (^{Ramachandran et al, 1995}). None of the 26 cases in this report included a Bungarus caeruleus bite and therefore a comparison cannot be made.

Daily renal potassium excretion is proportional to the dietary intake and an average adult looses about 80-120mmol of potassium each day (1-1.5mmol/kg/day) (^{Giebisch and Wang, 1996}; ^{Rabinowitz, 1996}). Urinary potassium excretion was extremely low in both patients, and neither patient excreted the 1-1.5mmol/kg/day (70 to 90mmol) of potassium given to them each day. This indicates near maximum renal conservation of potassium. Further, daily replacement of 1-1.5mmol/Kg of potassium did not normalize serum potassium in the deceased patient while it took three days in the survivor. This suggests that, extracellular potassium was either being lost to the gastrointestinal tract or being shifted intracellularly. The former is unlikely as none of the two patients had ileus or diarrhoea. Furthermore, the contribution of gastrointestinal loss in normal potassium homeostasis is minimal (^{Giebisch and Wang, 1996}). It could be argued that there was a significant intracellular shift of potassium in both patients similar to that seen in patients with barium poisoning (^{Bradberry and Vale, 1995}). As only 2% of total body potassium is extracellular and contributes to maintenance of resting membrane potential, shifting of potassium into the intracellular compartment should contribute significantly to neuromuscular weakness of these patients. The mechanism of this shift in Bungarus caeruleus is unclear and warrants further study.

The clinical features, EEG, VEP and BAEP, suggest that bungarotoxin induced deep and reversible coma is due to its effects on the cortical and brain stem structures. Hypokalaemia is most likely due to intracellular shifting of potassium. Knowledge of the potential underlying mechanisms of these phenomena could be used in the clinical management and would stimulate further research.

We thank the South Asian Clinical Toxicology Research Collaboration (SACTRC) for funding the laboratory studies and providing the digital EEG facility. SACTRC is funded by Wellcome Trust & Australian National Health and Medical Research Council International Collaborative Capacity Building Research Grant (GR071669MA ). IG is funded by AusAID grant no ALA 000379.