|Home | About | Journals | Submit | Contact Us | Français|
Study Objective: Data on poisoning with MCPA (4-chloro-2-methylphenoxyacetic acid) is limited to six case reports. Our objective is to describe outcomes from intentional self-poisoning with MCPA in a prospective case series of 181 patients presenting to hospitals in Sri Lanka.
Methods: Patient information was collected by on-site study doctors as part of an ongoing prospective cohort study of poisoned patients. History, clinical details and blood samples were obtained prospectively.
Results: Overall clinical toxicity was minimal in 85% of patients, including mild gastrointestinal symptoms in 44% of patients. More severe clinical signs of chlorophenoxy poisoning reported previously such as rhabdomyolysis, renal dysfunction and coma also occurred, but were uncommon. Eight patients died (4.4%). Most deaths occurred suddenly from cardiorespiratory arrest within 48 hours of poisoning; the pathophysiological mechanism of death was not apparent. The correlation between admission plasma MCPA concentration and clinical markers of severity of toxicity (physical signs, symptoms and elevated creatine kinase) was poor.
Conclusions: Intentional self-poisoning with MCPA generally causes mild toxicity, but cardiorespiratory arrest and death may occur. All patients should receive routine resuscitation and supportive care. It seems reasonable to correct acidosis and maintain an adequate urine output, but there is insufficient evidence to support other specific interventions. Our data do not support a clinical role for measurement of plasma MCPA in the acute management of poisoning and insufficient data were available to fully examine the utility of measured electrolytes and creatine kinase.
It is estimated that over 300,000 people die from pesticide poisoning each year in Asia and the Western Pacific.(1) Sri Lanka has a major problem with intentional self-poisoning, with high total and youth suicide rates.(2) The mortality from intentional self poisoning in Sri Lanka is high given ready access to potent pesticides by the largely rural population. Most poisoning deaths in Sri Lanka are due to organophosphorus (OP) pesticides, but deaths from intentional self-poisoning with other pesticides, such as chlorophenoxy herbicides are noted.(3) Toxicity from chlorophenoxy herbicides is also important in developed countries such as the United States where it is one of the most commonly used classes of pesticides.(4)
The most common chlorophenoxy herbicides are presented in Figure 1. The human literature on acute chlorophenoxy poisoning is limited to small case series and single reports of severe toxicity.(5) A third of these cases died and many others had clinical features of severe toxicity (Table 1). Most of these were reports of poisoning with 2,4-D. The selection of more serious cases for publication may mean that this does not represent the usual toxicity from chlorophenoxy poisoning nor an accurate estimate of their lethality. This is important given that some chlorophenoxy herbicides appear to have different mechanisms of toxicity and clinical features.(6-8)
MCPA (4-chloro-2-methylphenoxyacetic acid) is a widely used chlorophenoxy herbicide. It is the most important cause of chlorophenoxy poisoning in the North Central Province (NCP) of Sri Lanka.(3) There is limited information on clinical toxicity from MCPA poisoning, with only six single case reports in the literature.(9-14)
We describe the clinical outcomes of MCPA poisoning from data gathered during a large, prospective case series of patients who presented to three Sri Lankan hospitals with a history of intentional self-poisoning.
This was a prospective descriptive study of the exposure history and clinical details on consecutive patients admitted to three hospitals in Sri Lanka (Figure 2).
Patients presenting to three of the General Hospitals in Sri Lanka between 2nd April 2002 and 27th December 2003 with a history of MCPA poisoning were included. Resources are limited in these hospitals, particularly monitored beds in the ICU and wards and laboratory facilities. Blood tests are not usually obtained in most poisoned patients. All patients were stabilized in the ward and managed with supportive care. Patients with severe poisoning were admitted to ICU if beds were available.
All patients presenting to study hospitals with self poisoning were included. A large subset of these patients were part of a randomised controlled trial (RCT) evaluating the efficacy of activated charcoal (ISRCTN02920054). All patients with a history of acute poisoning were eligible for inclusion in this RCT, except for those who are under the age of 14 years, known to be pregnant or reported ingestion of either hydrocarbons alone or corrosives. Written informed consent was obtained prior to enrolment. Ethics approval was obtained from the University of Colombo, Sri Lanka and Oxford, United Kingdom.
Convenience blood samples were obtained on admission from 91 of the 113 patients recruited to the RCT between 8th May 2002 and 21st September 2003 for quantification of the plasma MCPA concentration using gas chromatography mass spectroscopy. These samples were in addition to those obtained by the treating physician for clinical indications.
In order to examine for the presence and time course of myotoxicity, serial convenience samples of blood were also obtained in 12 patients for measurement of creatine kinase (CK) activity. These patients presented to one hospital early in the series and were selected only because of the history of MCPA poisoning. CK activity was measured only in the 10 patients who tested positive to MCPA on admission. The medical chart of these patients was manually reviewed to correlate clinical features with plasma concentrations of MCPA and CK activity. Biochemical analyses were conducted by Queensland Health Scientific Services, Australia.
Treatment was determined by the treating physician and their team independent of the research team and irrespective of enrolment in the activated charcoal RCT. Supportive care with IV fluids, respiratory support, and inotropes were given where necessary, however alkalinisation was not used.
Important demographic details, including the type of poison and time since ingestion were collected on admission. Pre-determined demographic data and details of major clinical outcomes were prospectively recorded by on-site full time trained study doctors for every patient until discharge or death. These were entered directly into a hand-held computer data base at the time of recruitment and with each clinical review and evaluated by a frequency analysis. Other qualitative data were obtained directly from the medical chart for patients who died and those who provided serial blood samples.
A history of MCPA poisoning was reported in 181 patients, (145 of these patients were also recruited to the RCT of charcoal). Time to hospital admission and outcomes of poisoning are listed in Table 2. Ten patients received endotracheal intubation and ventilation for respiratory failure. The incidence of spontaneous vomiting was reviewed at one the study hospitals (accounting for the 63% of all admissions), and was noted to occur in 44% of patients.
Eight patients died (4.4%, Table 3). Seven of the eight patients died within 24 to 48 hours of poisoning from cardiorespiratory arrest of an unclear pathophysiological mechanism. A delay in the time to hospital admission did not appear to contribute to this outcome. One patient (Case 5) had bilateral crepitations and purulent secretions consistent with aspiration pneumonia. The medical chart could not be located for one patient (Case 8). Oliguria or dark-coloured urine was noted in five patients. Test bolus doses of atropine were given to some patients on admission who were initially suspected of being poisoned with anticholinesterase pesticides. This was generally not continued, but may have contributed to symptoms in some patients on admission.
Of the admission bloods available for analysis, MCPA was not detected in 11/91 patients (detection limit 0.05mg/kg plasma), consistent with minimal or no exposure to MCPA. One of these patients presented with a Glasgow Coma Score (GCS) of 15/15 six hours post-ingestion, but soon developed seizures and abdominal pain. The cause of these symptoms is unknown, but most likely represents poisoning with a substance other than MCPA, for example fipronil. The patient was discharged without complication four days later. All other patients with an undetectable MCPA level developed no signs of poisoning.
The relationship between admission plasma MCPA concentration, time since ingestion and severity of toxicity appears poor (Figure 3), although only one death (Case 6) was captured in this cohort, given the difficulty in obtaining consent from patients with an altered level of consciousness on presentation. Although CK activity tended to be increased in patients with higher admission MCPA concentrations (>300mg/kg), the correlation appears poor (Table 4 and Figure 4). With the exception of Cases B and C, CK activity was only marginally elevated in these patients. Case B presented with GCS 10/15 and profuse vomiting. Myalgias and weakness developed the following day and resolved within 48 hours. Case C presented with mild haematemesis, diarrhoea, abdominal pain and anuria despite fluid loading. Urine output returned by the next day and despite some ‘throat pain’, the gastrointestinal symptoms had settled. He complained of generalised myalgias, but appeared otherwise well and was discharged within 50 hours. No other investigations were ordered during this admission. The CK increased until discharge (Figure 4), which together with the myalgias, may be consistent with ongoing skeletal muscle toxicity.
This study was performed in secondary referral hospitals in rural areas of Sri Lanka. They have very limited resources and laboratory tests are not routinely performed. Monitored beds in the ICU or wards are not routinely available. The patient-doctor and patient-nurse ratios are high, so data on clinical features exceeding those obtained for the purpose of our study are limited. Most deaths occurred in unmonitored beds in the absence of blood tests, which limits our ability to determine the mechanism of death.. Further, as blood tests were only taken on those who consented to enrolment in the RCT, patients who were most unwell were less likely to have MCPA blood concentrations measured.
This is the first published prospective case series of major outcomes from chlorophenoxy self-poisoning. Mild gastrointestinal symptoms were noted, but overall, toxicity from MCPA poisoning was low and mortality was less than other common pesticide poisonings presenting to the same hospitals. Although animal data suggest that other chlorophenoxy herbicides are more toxic than MCPA (Figure 1), it seems unlikely that a third of poisonings with other chlorophenoxy herbicides will be fatal. The case series and single reports that this outcome was based on were probably published because of their severity or unusual presentation, representing a reporting bias.(5)
Variable clinical features have previously been reported for MCPA poisoning, but all are consistent with those described in our patients (Table 1). We found that 85% of patients had minimal toxicity (despite a large reported exposure in some) and mortality was 4.4%. This in-hospital death rate is higher than that noted from medications, but lower than that of other pesticides including the herbicides glyphosate (5%) and paraquat (60%) and organophosphorus insecticides (11%).(15) Blood levels confirmed ingestion of MCPA in 88% of samples tested and there was only one patient where the combination of laboratory investigations and clinical features suggested that MCPA was not responsible for the clinical presentation despite a history of ingestion. Therefore these data should be an accurate reflection of the clinical course from acute intentional self poisoning with MCPA.
Other clinical features of MCPA poisoning were not systematically recorded in each patient, but abdominal pain, myalgia, fasciculations, throat pain and miosis were noted by study doctors in some patients.
Seven of eight deaths were due to a cardiorespiratory arrest of an unclear aetiology. Speculation of factors leading to death in these patients is restricted by the very limited availability of laboratory tests. Oliguria and discoloured urine was noted in five of these patients, which could indicate renal toxicity and be an important predictor of severe toxicity or death.
Uncoupling of oxidative phosphorylation may be an important component in the cause of death in patients with large chlorophenoxy herbicide exposures. Chlorophenoxy herbicides uncouple oxidative phosphorylation in vitro through an unclear mechanism of action.(16;17)
Uncoupling describes the process whereby oxygen consumption and heat production increase out of proportion to the generation of adenosine triphosphate.(18) It may be caused by extrinsic factors such as chemicals or drugs which disrupt mitochondrial function.(19) Initially this leads to an increase in mitochondrial respiration,(20;21) until the declining levels of ATP are insufficient for essential cellular functions including active transport pumps such as Na-K ATPase. This leads to loss of cellular ionic and volume regulation, and if ATP is not supplied promptly, the effect is irreversible and cell death occurs.(19)
Clinical features consistent with marked uncoupling of oxidative phosphorylation have been reported in cases of human chlorophenoxy herbicide poisoning. Tachypnoea with respiratory alkalosis have been noted in cases of severe poisoning, some of whom died (5;13)(D Roberts and P. Piyasena, unpublished observations), which may be consistent with increased mitochondrial respiration from mild uncoupling. More severe poisoning may be characterised by severe hypoxia, metabolic acidosis, hyperventilation, hyperkalaemia, fever, elevated creatine kinase, tachycardia, generalised muscle rigidity, hypotension, and cardiac asystole.(5;7;22;23) Some of these features were noted in patients who died in our series (Table 3, patients 3,4,7, and perhaps 8). However, these data were not in the dataset collected prospectively under the protocol.
The relationship between plasma chlorophenoxy herbicide level and toxicity is not clear. A depressed level of consciousness has been reported with plasma chlorophenoxy concentrations from 80 mg/L to over 1000 mg/L,(24) while concentrations more than 500 mg/L may be associated with severe toxicity.(25) But death occurred in two patients with a blood MCPA level of 180 mg/L and 230 mg/L.(11;13) A patient survived severe MCPA poisoning (hypotension and limb myotonia) despite a blood concentration of 546 mg/L two hours after ingestion. The muscle effects resolved after a number of days when the MCPA level was less than 100mg/L.(9)
The discordance between admission MCPA levels and the peak toxicity observed during hospital admission is noted in Figure 3. In contrast, concentration dependent uncoupling of oxidative phosphorylation has been demonstrated with rat mitochondria in vitro.(16) The poor correlation between clinical toxicity and plasma concentrations may reflect poor correlation between plasma (measured) and intracellular (mitochondrial) concentrations. Mechanisms which alter the distribution of MCPA between these two spaces includes MCPA-induced damage to cell membranes which increases penetration of the herbicide through cell membranes.(5;26-30) In rats, high plasma concentrations of MCPA damage cell membranes and induce toxicity, but the correlation between plasma levels, membrane damage and toxicity is poor.(29;30)
Another potential mechanism is variation in blood pH, which alters protein binding and tissue distribution. Acidosis increases the proportion of non-ionised MCPA (pKa 3.07), which is lipophilic, and therefore readily crosses the cell membrane, increasing the intracellular concentration. This has been observed in in vitro studies and is similar to that observed for aspirin. (23;31) Similarly, an alkaline plasma pH is expected to decrease tissue binding and increase plasma levels.(25;32;33) Therefore, the distribution of MCPA between plasma and intracellular spaces is likely to vary with plasma pH.
Routine measurement of plasma MCPA levels in patients with intentional self poisoning does not appear useful. Instead, quantification of markers of intracellular toxicity, such as CK from skeletal muscle, may be useful to monitor the effects of MCPA poisoning. Of the patients who had serial blood samples, only two patients (Cases B and C) had significant toxicity (Table 4, Figure 4). Myalgias were reported by these two patients, particularly in Case C where the CK was 1030 on admission and increased until discharge. In most patients, peak CK activity was delayed (Figure 4).
All patients should receive routine resuscitation and supportive care. There is insufficient evidence to describe other specific managements of acute MCPA poisoning. It would be reasonable to correct acidosis based on the likely adverse effect of acidosis on MCPA distribution kinetics leading to increased intracellular concentrations.
All significant poisonings (ie symptomatic oral ingestions) should be treated cautiously, preferably in an ICU or other unit with continuous monitoring for 24 to 48 hours. In our series, initial mild toxicity as indicated by normal vital signs and level of consciousness on admission did not preclude subsequent severe toxicity and death. However, we observed that no patients who were completely asymptomatic for the first six hours went on to a fatal outcome. Late deterioration in a reasonably well patient has also been reported with poisoning with MCPP (a related chlorophenoxy herbicide). In this case severe metabolic derangement and death occurred in a patient who was fully conscious prior to an asystolic arrest.(7)
Our data suggest measurement of plasma MCPA confirms exposure but does not provide further information on the severity of poisoning. Confirmation of MCPA exposure may not be necessary in patients presenting with intentional MCPA self-poisoning given that the history was accurate in 88% of patients in our series. We had insufficient data to suggest whether electrolytes, blood gases or CK may be better indicators of the severity of poisoning.
An adequate urine output (>1mL/kg) should be ensured in all patients. Urinary alkalinisation (pH >7.5) and an adequate urine output limits reabsorption and promotes renal excretion.(23;33-35) It might be reasonable to consider these measures in patients who are symptomatic, particularly if they have signs of toxicity. A randomised controlled trial is required to determine the efficacy of this treatment.
Patients with severe toxicity require prompt and effective plasma and urinary alkalinisation, and possibly haemodialysis if facilities are available.(5;35) While there is limited information to support alkalinisation, it is rarely associated with adverse effects when administered carefully and with close observation.(35;36) Signs consistent with uncoupling of oxidative phosphorylation are likely to be associated with a poor outcome.
A better description of the clinical, biochemical, histopathological and histochemical features of severe toxicity and death is required to confirm the mechanism of death in patients. This may help to guide clinical management.
We found that toxicity from MCPA poisoning is low compared to some other agrochemicals, with 85% of patients having minimal clinical toxicity; but toxicity is still significant with a mortality of 4.4%. Further research is required to identify early markers of poor outcome and to determine if alkalinisation or other interventions can improve the prognosis in this group.
Study doctors who assisted with data and sample collection, and hospital doctors for their cooperation with the study. Manel Abeyawardene, Vincent Alberts Brock Jones, and in particular Mary Hodge from Queensland Health Pathology and Scientific Services for assistance with analyses. The Medical Superintendent of Anuradhapura General Hospital for permission to access medical records, and Medical Records staff who located these records. Jason and Alison Roberts for assistance with locating some of the older references.