Organophosphates (OP) inhibit acetylcholinesterase and cause excessive acetylcholine accumulation, which affects muscarinic and nicotinic receptors at synapses within the peripheral and central nervous systems [1]. OP poisoning causes neurotoxic sequela and has a high mortality rate [2-4].

The World Health Organization (WHO) estimates that the incidence of pesticide poisoning in developing countries has doubled during the past 10 years [5]. In Korea, the mortality due to poisoning has increased rapidly since 1998 [6]. Herbicides and pesticides constituted the largest proportion of deaths due to poisoning of all causes, although OP were not evaluated specifically [6].

Combined atropine and pralidoxime therapy is the cornerstone of OP poisoning treatment. Extracorporeal elimination may be a useful adjunctive strategy. However, the use of hemoperfusion in the management of severe OP poisoning remains controversial [7-10].

Recently, a difference among OP insecticides in human self-poisoning was reported. Moreover, the toxicity differed in humans and animals. These finding suggest that each OP has a different toxicity, although textbooks and articles have traditionally considered acute OP poisoning as a homogenous entity.

This study determined the predictors of a poor outcome in patients with acute OP poisoning and differences of toxicity according to OP type.

The general characteristics of the 68 study subjects, who all had acute OP poisoning at presentation, are summarized in Table 1. The mean amount of ingested OP was 60 mL (range, 5 to 500) and the agents responsible for poisoning were dichlorvos (n=21), O-ethyl-O-4-nitrophenyl phenylphosphonothioate (EPN, n=12), phosphamidon (n=8), phenthoate (n=4), malathion (n=3), methidathion (n=2), profenofos (n=2), chlorpyrifos (n=1), and an unidentified OP (n=15).

The RF group included 35 of the 68 patients. The mean duration of mechanical ventilation was 9.5±8.3 days. A tracheostomy was performed in 17 of these 35 patients. Table 2 shows the comparison of initial parameters in the RF and non-RF groups. The APACHE II score, white blood cell count, and lactate dehydrogenase and aspartate aminotransferase levels were higher in the RF group, and the serum cholinesterase and bicarbonate levels were lower than the non-RF group. No significant difference in the arterial oxygen partial pressures at presentation was observed between the two groups. The APACHE II score was a significant predictor of respiratory failure (odds ratio [OR], 1.273; p<0.01; 95% confidence interval [CI], 1.122 to 1.444). In addition, the APACHE II score was an important predictor of RF in the multiple logistic regression analysis (Table 3).

Information on the patients according to hemoperfusion treatment is presented in Fig. 1. The APACHE II scores were higher in patients with hemoperfusion treatment, but hemoperfusion treatment did not affect mortality. Regardless of hemoperfusion treatment, the APACHE II scores were higher and cholinesterase levels were lower in those who died.

Thirteen of the 68 patients died. The clinical parameters are summarized in Table 4. The serum cholinesterase levels were lower in those who died. Twelve of the 35 RF group patients and only one of the non-RF patients died. Therefore, the need for mechanical ventilation was a significant predictor of a poor outcome (OR, 16.70; p<0.01; 95% CI, 2.03 to 137.60).

In addition, the APACHE II score was a significant predictor of mortality (OR, 1.194; p<0.01; 95% CI, 1.089 to 1.309). An APACHE II score <10 was a significant predictor of mortality (OR, 16.11; p<0.01; 95% CI, 3.17 to 81.73); a serum cholinesterase <1,000 IU/L was significantly associated with a poor outcome (OR, 0.676; p<0.05; 95% CI, 0.54 to 0.845).

Pearson's correlation analysis revealed a significant negative correlation between the APACHE II score and serum cholinesterase level (r=0.415, p<0.01, Fig. 2).

An age over 60 years was also a significantly predictor of a poor outcome (OR, 6.852; p<0.01; 95% CI, 1.677 to 28.003). Hemoperfusion treatment was not related to a better outcome, although the patients who underwent hemoperfusion were in poorer condition (e.g., their APACHE II scores were higher).

Regarding the agents responsible, the death rate was 0% (0 of 21 of cases) for dichlorvos, 16.7% (2/12) for EPN, 37.5% (3/8) for phosphamidon, 25% (1/4) for phenthoate, 0% (0/3) for malathion, 50% (1/2) for methidathion, 0% (0/2) for profenofos, 0% (0/1) for chlorpyrifos, and 40% (6/15) for unidentified OP.

OP poisoning is a serious clinical entity and causes considerable mortality and polyneuropathy in survivors that is not always completely reversible. The estimated mortality following OP ingestion ranges from 20 to 50% [3,10-12].

In this study, the total mortality was 19% (13 deaths in 68 patients), and 12 of these deaths occurred in the 35 patients with respiratory failure. In our study, age, amount ingested, APACHE II score, initial cholinesterase level, and respiratory failure requiring mechanical ventilation were significantly associated with a poor outcome. The usefulness of the serum cholinesterase level in this context remains controversial [13,14]. In our patients, the cholinesterase level was correlated with the APACHE II score (Fig. 2), but this findings casts doubt on the usefulness of the cholinesterase level in patients with a low APACHE II score. The APACHE II score is a reliable general index that is useful for evaluating a wide spectrum of patients in intensive care units [15], and essentially shows that the degree of physiological derangement is closely correlated with outcome in critically ill patients. In a previous study, an APACHE II score >26 was a poor prognostic indicator in cases of OP poisoning [16], and the both APACHE II score and Glasgow coma scale predicted outcome [17,18]. In our study, the APACHE II score was correlated with respiratory failure and outcome. In particular, an APACHE II score >10 predicted a poor outcome.

Extracorporeal methods of blood purification have recently been recommended as a therapeutic approach to OP poisoning because of their ability to remove xenobiotics from blood, particularly hemoperfusion systems involving activated charcoal or resins [9,19,20]. However, the therapeutic benefits of this approach have not been substantiated by other investigators, and this has been explained by the large volume of distribution. The use of hemoperfusion for managing severe OP poisoning is also controversial [7,9]. In a study of 52 OP-poisoned patients, Altinop et al. [7] reported that hemoperfusion was useful in severe cases. By contrast, in our study, hemoperfusion was ineffective for improving survival, although it should be emphasized that the patients who were hemoperfused had high APACHE II scores, which makes it difficult to evaluate the effect of hemoperfusion. Randomized controlled studies are required to investigate the effect of hemoperfusion on survival in patients with severe OP poisoning.

Various types of OP were included in this study. No death occurred in 21 patients with dichlorvos poisoning, although the mean patient APACHE II score was 5.52±5.01. By contrast, 2 of 12 EPN cases, 3 of 8 phosphamidon cases, and 1 of 4 phenthoate cases died. This suggests that different OP have different toxicities. Eddlestone reported differences in the toxicity of OP [21]. In addition, the human toxicity differed from the animal toxicity suggested by WHO methods based on the toxicity in rats after oral dosing. That report suggested that each OP be considered as an individual poison. Our findings were consistent with that opinion.

We conclude that the initial APACHE II score is a useful prognostic indicator in cases of OP poisoning. Each OP might have a different toxicity. The management protocol for each OP in humans needs further study. Moreover, our results suggest that hemoperfusion is of limited benefit in cases of severe OP poisoning. Further randomized controlled studies of the beneficial effects of hemoperfusion on patient survival are required.