Neuropsychological and Psychiatric Functioning in Sheep Farmers Exposed to

Organophosphate Pesticides

DEFRA Project VM02302

S MACKENZIE ROSS (Principal Investigator)

V HARRISON (Research Assistant)

K ABRAHAM (Senior Research Assistant)

T HUGHES (Research Assistant)

J BRITTON (Research Assistant)

V CURRAN (Co-Applicant)

C BREWIN (Co-Applicant)

Sarah Mackenzie Ross, DPsychol, is a Consultant Clinical Neuropsychologist, Val Curran is

professor of psychopharmacology and Chris Brewin is professor of clinical psychology, in

the Research Department of Clinical Educational & Health Psychology, University College

London. Virginia Harrison, MSc, Kelly Abraham, BSc, Tessa Hughes, BSc and Julia Britton,

MSc were all Research Assistants who worked on this project at some time over the last 4

years.

Address for correspondence: Dr S Mackenzie Ross, Sub-department of Clinical Health

Psychology, University College London, Gower Street, London WC1E 6BT, United

Kingdom.

KEYWORDS

ORGANOPHOSPHATES

PESTICIDES

COGNITIVE IMPAIRMENT , MEMORY, MOOD

1

EXECUTIVE SUMMARY

ObjectivesTo establish whether low level exposure to organophosphate pesticides is associated with neuropsychological impairment and psychiatric disorder in UK sheep farmers; to determine the nature and extent of neurobehavioural problems following low level exposure to organophosphate pesticides; to investigate whether some individuals are more vulnerable to the effects of OPs than others.

MethodA cross-sectional study in which the performance on neuropsychological tests of 132 sheep farmers with a history of low level exposure to organophosphate pesticides (insufficient to produce acute intoxication) was compared with 79 non-exposed healthy volunteers (police workers) matched for age, gender, years in education and level of intelligence. In order to take account of the ‘healthy worker’ effect, both working and retired farmers and controls were included in this study. Rigorous exclusion and inclusion criteria were used in this study which meant that over 60% of exposed and unexposed individuals identified as potential study participants were subsequently excluded.

All participants underwent detailed neuropsychological testing. Well known, standardized and clinically sensitive tests were used which are routinely used in NHS services for diagnostic purposes. Sheep farmers were interviewed about their work and exposure history. All participants completed a questionnaire regarding their physical health. Mood state was examined by a structured clinical interview (SCID) and questionnaire measures.

Genetic differences between individuals render some people more susceptible to the toxic effects of certain chemicals than others. For example, the human paraoxonase 1/arylesterase enzyme (PON1) plays an important role in the detoxification of organophosphates and helps protect against the potentially harmful effects of OPs. Each study participant was asked to provide a sample of blood for determination of PON1 status.

Results: A range of emotional, physical and cognitive problems were identified in agricultural workers with a history of low level exposure to OPs. In terms of cognitive function, general intellectual ability, reasoning, visuo-spatial and verbal ability were relatively well preserved, but agricultural workers obtained lower scores on tests of response speed, working, verbal and visual memory, mental flexibility and fine motor control, than non-exposed controls. These differences remained after controlling for Type 1 errors, depression, removing participants with a history of ‘dippers flu’; and irrespective of whether exposed farmers were compared to rural police workers or with published test norms derived from a cross section of several thousand adults in the general population. Therefore, these findings are unlikely to have occurred by chance or to be due to confounding factors such as mood, acute exposure or selection of an inappropriate control group.

As far as we are aware, this is the first study in the UK to take account of the healthy worker effect and include individuals who have retired on ill health grounds. Although higher rates of emotional distress and physical symptoms were reported by retired farmers few differences were found on objective measures of cognitive function or potential vulnerability factors such as PON1 status. Individuals who have retired on ill health grounds do not appear to be at increased risk of suffering cognitive impairment following exposure to OPs.

A number of significant correlations were observed between duration of exposure and verbal and visual memory, verbal ability, strategy making and fine motor control. Although weak, they were in the expected direction, consistent with findings from the group analyses and consistent with study hypotheses. Binomial tests suggest they are unlikely to have occurred by chance.

2

ConclusionsBoth correlation and group analyses suggest a relationship may exist between low level exposure to organophosphates and impaired neurobehavioural functioning. The cognitive deficits identified in this cohort can not be attributed to mood disorder, malingering or poor effort on testing, a history of acute exposure, a past medical or psychiatric history that could otherwise account for ill health, genetic vulnerability in terms of PON1 polymorphisms, chance or because an inappropriate control group was selected. The cohort of farmers included in this study were relatively fit, none were poor metabolisers of OPs and yet they show evidence of neurobehavioural impairment. The pattern of deficits identified in this study is consistent with reports from previous studies and consistent with what would be expected given the principle action of OPs (i.e. inhibition of acetylcholinesterase) and the distribution of cholinergic cell groups in the brain.

ImplicationsThe results of this study suggest there may be a relationship between long-term low-level exposure to organophosphates and the development of neurobehavioural problems. This has implications for working practice and policies and guidelines about the use of organophosphate chemicals on the farm should be reviewed.

Follow-up studies should be carried out to determine whether symptoms persist over time, improve or worsen. At present, there are no recommended treatment protocols for individuals who report chronic ill health following exposure to OPs, so there is a need for prospective treatment trials.

It is also important to consider the possibility that clear cut dose-response relationships that might be discernable following acute exposure may not be apparent with low level exposure. Low level exposure may produce subclinical neurological injury that accumulates over time and only becomes apparent when specialised neuropsychological or neurological tests are used to evaluate patients or when neuronal reserves are depleted by processes such as ageing, thus unmasking deficits.

3

INTRODUCTIONOrganophosphate pesticides (OPs) are being increasingly used around the world for a variety of agricultural, industrial and domestic purposes. Concerns have been expressed about the effects of these chemicals on human health, but there is a lack of reliable data on the scale of the problem. The immediate effects of high-level exposure to OPs have been well documented and involve inhibition of the enzyme acetylcholinesterase, causing changes in peripheral, autonomic and central nervous system function (cholinergic crisis).1

The possibility that long-term low-level exposure to OPs in doses below that causing acute toxicity may cause ill health is controversial. In 1998 a working party was commissioned by the UK Government to review the available scientific evidence concerning the potential toxicity of low level exposure to OPs. The Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment (COT) considered clinical data (i.e. individual case reports) and reviewed published scientific papers on this subject. They concluded that although evidence exists to support the view that high level / acute OP poisoning can cause ill health, further research was needed to determine whether low level exposure to OPs causes disabling disease as previous research had yielded inconsistent results.1-7

In 2000 the UK Government agreed to fund a programme of research to address topics identified for further research by this working party. One of COT’s recommendations was that any study designed to investigate the chronic effects of OPs on human health, must take account of the ‘healthy worker effect’. The healthy worker effect is a phenomenon observed initially in studies of occupational diseases, whereby workers have lower morbidity and mortality rates than the general population because those with disabling levels of disease are either excluded from employment or unable to continue in employment8.

Most of the previous research looking at the long term health effects of exposure to OP’s has examined individuals who are fit enough to be in employment at the time of investigation and have not allowed for the fact that individuals with disabling disease may have retired from work or work in a different capacity. Therefore, the risk associated with exposure to OPs may have been underestimated because of this common selection bias.

It is also possible that some individuals are more vulnerable to the toxic effects of OPs than others because of their genetic make up, exposure, psychosocial, or medical history. COT recommended future studies explore factors that may render some individuals at increased risk of clinically significant disease rather than simply looking for effects on the mean level of quantitative health indices in any exposed population.

The ultimate aim of the present study is to establish whether low level exposure to OPs is associated with disabling neuropsychological and psychiatric disease in a small subgroup of farm workers. The occupational group examined in this study were sheep farmers, as organophosphate pesticides were used extensively in the dipping of sheep in the UK and farmers are generally considered to have relatively low-level exposure to OPs. A further aim of the study was to determine whether individuals who have retired on ill health grounds constitute a particular subgroup of individuals who are more susceptible to the effects of OPs than others.

The current study will examine the hypothesis that study participants will show a similar pattern of cognitive and emotional deficits as that reported in earlier studies of individuals with a history of low level exposure to OPs. They will perform more poorly on tests of response speed, working and general memory, mental flexibility and they will have higher rates of emotional distress. Deficits in perceptual, visuo-spatial, intellectual reasoning and general verbal abilities are not expected1.

4

METHOD

Scientific Objectives

Objective 1: To establish whether farm workers with a history of low level exposure to OPs (insufficient to cause acute intoxication) show evidence of physical disease, cognitive impairment and / or mood disorder.

Objective 2: To determine the nature and severity of physical symptoms, neuropsychological abnormalities and psychiatric disorder in farm workers with a history of low level exposure to OPs.

Objective 3: To investigate whether background factors (e.g. psychological profile, PON1 status, medical or exposure history) render some individuals vulnerable to the effects of OPs.

Ethical Approval

Ethical Approval for this study was granted by the joint University College London / University College London Hospital committee and written informed consent was obtained from all study participants. Study Participants were aware of the overall purpose of the study i.e. to determine whether exposure to hazardous chemicals on the farm has any adverse effects on health. However they were not aware of the specific hypotheses of this study concerning the profile of deficits and retained abilities expected.

Study Design

This is a cross sectional, case control study in which performance on neuropsychological tests of working and retired farmers, exposed to OPs in the course of their work, was compared with non exposed, working and retired healthy controls. It is a 2x2 design which essentially means two studies have been undertaken:

Study 1: A cross-sectional study in which groups of exposed and unexposed individuals are compared.

Study 2: A case control study in which working and retired farmers are compared. Cases are farmers who have retired on ill health grounds and controls are working farmers.

Statistical PowerFinding comparable studies in order to calculate power calculations proved difficult. As was mentioned in Chapter Three, studies have utilized different methodological designs, populations, psychometric tests and control groups making direct comparisons difficult. Studies by Stephens (1995) and Rosenstock (1991)9,10 found moderate effect sizes (0 .51 -0 .59) between cognitive function and exposure history indicating that future studies would require a sample size of sixty two individuals to have 80% power to detect a relationship of this magnitude between neuropsychological functioning and exposure history. Power calculations based on previous work by Mackenzie Ross et al (2007)27 of farmers who have retired on ill health grounds suggest a sample size of 40 farmers/controls would be sufficient to detect a relationship between neuropsychological functioning and exposure history.

Participants

(1) Exposed cohort:

Two groups of sheep farmers were recruited. 79 working sheep farmers and 65 sheep farmers who had retired on ill health grounds. Both groups had a history of low level exposure to OPs, insufficient to cause acute intoxication resulting in medical intervention. Prior acute exposure was assessed by interview at the recruitment stage and again during the clinical study. Participants were asked whether they had ever felt so unwell immediately after dipping that they sought medical advice/intervention within 48 hours. If they had, they were excluded from the study. Remaining participants were asked if

5

they had ever suffered symptoms of ‘dippers flu’ after dipping for which they had not sought medical intervention.

Inclusion Criteria for Exposed Cohort:

Aged between 18-70 years old. Exposure to organophosphate pesticides for a minimum of 5 years prior to 1991 (safety

regulations were implemented in 1992). NO history of acute intoxication requiring medical intervention. Living in the South West or North of England. For the retired cohort, they must have retired on ill health grounds NOT age or economic

reasons.

Since the main aim of this study was to determine whether there is a relationship between low level exposure to OPs and cognitive impairment, it was important to exclude any participants with a medical or psychiatric condition which might otherwise account for any deficits identified during assessment.

Exclusion Criteria for Exposed Cohort:

Those with a history of acute organophosphate intoxication Substance abuse (including alcohol) History of psychiatric problems prior to exposure, neurological or serious medical problems

which might otherwise account for any cognitive or emotional problems identified in the study - refer to Table 2.

(2) Healthy Controls:

It is extremely difficult to find a group of farmers in the UK who do not have a history of exposure to OPs and so it was necessary to identify a different occupational group to act as controls. A number of occupational groups were considered, but a primary concern was to find an occupational group with sufficient numbers of individuals who have retired on ill health grounds, that could be easily identified and accessed. Other important criteria included:

The control group should be matched to farmers in terms of characteristics which have been shown by previous research to affect cognitive function (the main outcome variable of this study) i.e. age, gender, education level, premorbid IQ.

The control group should not have a history of exposure to organophosphates.

Variables such as the exact nature of the work undertaken, location, lifestyle, attitudes, life experiences were considered to be less important since these variables have not been shown by research to significantly affect performance on psychometric tests.

Rural police workers who have never worked in the farming industry were recruited as controls. Rural police workers undertake both administration and outdoor work as do farmers and a major advantage of using the police as a control group is that an organisation exists which holds a database of 80,000 retired members of the police force. Furthermore, the police force is divided into local constabularies making it possible to recruit police workers from the same geographical regions as sheep farmers.

None of the police workers included in the study had a history of exposure to OPs. 40 rural police workers who were fit enough to be in employment and 42 rural police workers who had retired on ill health grounds were examined.

6

Inclusion Criteria for Healthy Controls:

Aged between 18-70 years old Retired on ill health grounds NOT age or economic reasons No exposure to organophosphate pesticides Has worked in a rural area in the South West or North of England

Exclusion Criteria for Healthy Controls

Exposure to organophosphates A career that has been metropolitan Substance abuse (including alcohol) History of psychiatric, neurological or serious medical problems which might otherwise

account for any cognitive or emotional problems identified in the study.

Recruitment

Identification of study participants: Sheep farmers.

For practical reasons, the focus of the project was restricted to two geographical areas of England with the highest number of sheep, according to DEFRA’s ‘Distribution of Sheep in UK on 02 June 2005’. These two areas were the North and South West of England.

This study sought to identify from these regions, working and retired farm workers / farmers with a history of low level exposure to OPs. Three methods of sampling were used to identify the target population.

(1) Purposive sampling – written correspondence(2) Purposive sampling – telephone contact(3) Advertising

Purposive sampling – written correspondence

Contact details of farm owners in the south west and north of England were purchased from databases held by (1) a company called Experian which owns the right to sell data from the UK National Business Directory (2) a company called Tri-Direct which owns the right to sell the membership lists of the National Farmers Union (3) The Royal Agricultural Benevolent Institution (RABI) which provides welfare advice to working and retired farmers in need, especially those who are elderly or disabled. These companies restricted the amount of contact with their members. Members could be contacted on only one occasion, by letter. It was not possible to telephone members.

Letters were sent to all of the farm owners on these databases asking them to provide us with the contact details of any sheep farmers who had retired on ill health grounds that were known to them. They were offered a small financial incentive for any contacts that subsequently met our inclusion / exclusion criteria. Eight thousand, two hundred and sixty two farm owners were contacted. The response rate was poor and less than 2% for business directories and 4.5% for RABI. Nominees identified from this method of sampling were subsequently contacted by telephone and interviewed to establish some basic facts about them including their reasons for retirement and exposure history.

Purposive sampling – telephone contact

The contact details of sheep farmers in the south west and north of England were obtained from the Wool Marketing Board (WMB). The WMB gave us this information free of charge as they were keen to assist us with this study. Over fifteen thousand farmers were listed on this database, twelve thousand of whom were new to us in the sense that they were not included on the other databases

7

mentioned in this report. We contacted every fifth person on the WMB database, up to a total of three hundred and ninety three farmers, by telephone and explained the purpose of this study, asked them if they had retired on ill health grounds or could suggest any other farmers who might have retired on ill health grounds. The response rate from this method of sampling was much greater than written correspondence at 59%.

Advertising

Details about this study were published in farming newspapers and publications, in organisations’ newsletters such as union newsletters (NFU & TGWU) and support organisation newsletters (OPIN, PAN, RABI). Information about the study was also distributed at agricultural shows and sent to a number of rural GP surgeries. The study was also described in several regional radio broadcasts (circa 17) and on the Farming Today programme (twice) between 2005 and 2007.

Identification of study participants: Police workers.

The police force is divided into local constabularies and the Human Resources and Occupational Health Departments for each of the 12 regions we had recruited agricultural workers from were contacted and their assistance was sought in recruiting working and retired police workers into the study.

The National Association of Retired Police Officers (NARPO), which holds a database of 80,000 retired members of the police force, assisted us in recruiting rural police workers who had retired on ill health grounds. Two police convalescence and treatment centres in the UK, the Northern Police Convalescent and Treatment Centre in Harrogate and the Police Convalescent Home in Berkshire, also assisted us in recruiting rural police workers who had retired on ill health grounds.

Details of the study were emailed by police constabularies and NARPO to police workers on their database and the study was advertised in Police Press and associated websites (local police magazines and newsletters and national publications such as Police Life, Police Oracle, Police Review). Posters advertising the study were also placed in a few local police stations.

Measures

(1) Exposure HistoryEach sheep farmer underwent a semi-structured interview about their work and exposure history (see appendix 1). Individuals were asked to specify when they began working with OPs, in what capacity, their level of exposure in terms of frequency and duration, the use of protective clothing, their involvement in high risk activities such as diluting concentrate, their use of other agricultural chemicals, the onset of their physical/psychological problems and the temporal relationship with exposure to OPs, and whether or not they had a history of acute poisoning (i.e. ‘dippers flu’) and whether or not they had felt so unwell after dipping that they sought medical help.

In addition to obtaining information about specific aspects of exposure, two exposure indexes were calculated for each participant based on (1) the number of days per year spent using OPs multiplied by the number of years spent using OPs (lifetime exposure index) (2) the concentration of the chemical in the skin contamination layer, the area of skin contaminated (both estimated from job title) and duration of exposure (Esk exposure metric). This Exposure Metric was used in the SHAPE study11, 12.

Rural police workers were only included in the study if they did not have a history of exposure to potentially toxic chemicals. This was established by telephone interview at the recruitment stage of the study. They were asked if at any time in their lives they had been exposed to or worked with potentially toxic chemicals and if they had ever felt unwell after being exposed to chemicals. Thirty three police workers had to be excluded because they had assisted farmers with sheep dipping in the past.

8

(2) Cognitive AssessmentAll participants (exposed and control cohorts) underwent neuropsychological assessment. Well known, standardized and clinically sensitive tests which are commonly used in routine clinical practice within the National Health Service were selected. All tests had adequate published reliability, validity and normative data. The Researchers who evaluated study participants were aware of which group they came from (e.g farmer or police worker) but were blind to the participants’ exposure history whilst assessing their cognitive function and mood state.

Intellectual ability - the Wechsler Adult Intelligence Scale-III was administered to assess current intellectual functioning13. This test is composed of fourteen subtests. Seven measure verbal skills and are predominantly, but not entirely associated with the functioning of the left hemisphere of the brain. Seven measure non-verbal skills and are associated with right hemisphere functioning. Subtests measure a range of cognitive functions such as general knowledge, vocabulary, arithmetic, verbal and visual reasoning, working memory, response speed and visuo-spatial ability. Two of the subtests are optional. These were not administered.

Memory - The Wechsler Memory Scale – III (short version) was used to assess working, visual and verbal memory14. The following subtests were administered, logical memory (story recall), verbal paired associates, letter/number sequencing, digit span, spatial span, face recognition and family pictures.

Response speed and mental flexibility were assessed by a number of means (1) Trail Making Tests A and B15 (2) The California Computer Assessment Package (CALCAP) which measures simple and choice reaction time16. The Stroop test was included as a measure of mental flexibility17.

Language - the Graded Naming Test was administered to assess naming ability18. A verbal fluency test (FAS) was used to assess expressive language19.

Fine Motor Skill – the Grooved Pegboard was used to assess motor dexterity20. This test is very sensitive to general slowing due to medication, toxic effects and diffuse brain injury.

Effort / Malingering – A brief computerised verbal memory screening test which measures a person’s effort on testing, was included in the battery to ensure psychometric test results were valid. It is insensitive to all but the most extreme forms of cognitive impairment whilst being very sensitive to poor effort and exaggeration of cognitive difficulties21.

(3) Mood State

Mood State and Life Events – Mood state was assessed by two means: (1) A structured clinical interview (SCID)22 and (2) Questionnaire measures such as The Hospital Anxiety and Depression Scale23 and the Beck Anxiety24 and Depression Inventories25. A life events checklist26 was included to act as a prompt when interviewing participants about recent stressful life events.

(4) Physical Health

A questionnaire concerning physical symptoms was constructed (see appendix 2). It was based on a number of existing questionnaires with known reliability and validity: (1) the SF36 Health Survey, a multi-purpose, generic measure of disease burden (2) the Illness Perception Questionnaire-Revised. (3) the symptom checklist was examined to ensure it included symptoms, identified by previous research as being associated with exposure to pesticides. Additional symptoms which are not associated with exposure to OPs, were added to the list such as hearing loss, toothache, tinnitus, hay fever. The questionnaire was not subject to any form of validation during this study.

9

Participants were asked to give an overall rating of their health and to state the degree to which their work and social life had been affected by poor health. The questionnaire then provides a list of 39 symptoms and participants were asked to state whether they suffered from these symptoms, if so, when the symptom first appeared, the frequency with which the symptom is experienced, symptom severity, level of distress caused and the degree to which the symptom interferes with daily activities.

(5) Possible genetic vulnerability factors - Blood Analyses (determination of PON1 status)

Genetic differences between individuals render some people more susceptible to the toxic effects of certain chemicals than others. For example, the human paraoxonase 1/arylesterase enzyme (PON1) plays an important role in the detoxification of organophosphates and helps protect against the potentially harmful effects of OPs1.

Each study participant was asked to provide a sample of blood from a vein in their arm (separate consent was obtained for this aspect of the study). This was then centrifuged and the plasma removed into a cryogenic vial and frozen. Plasma samples were later sent to Professor Clement Furlong’s laboratory, University of Washington, Seattle, USA, for determination of PON1 status.

PON1 status was determined under conditions that allow for determination of in vivo rates of diazoxon detoxification (Richter and Furlong, 1999)34. An individual’s PON1 status was initially established by plotting rates of diazoxon hydrolysis at high NaCl vs. rates of paraoxon hydrolysis also at high salt. Further research by Furlong’s laboratory has compared these original assay conditions with the diazoxonase assay run under physiological conditions and found the data points from the two determinations are highly correlated. Conversion factors have since been developed for determining in vivo rates of diazoxon hydrolysis from other assays (Richter et al. 2008 and 2009)34b,34c. Please see Appendix 10 for further details of PON1 assay conditions and procedures for status determination.

10

RESULTS

Overview

Section 1, presents the number of participants identified by different sampling methods followed by demographic information regarding the participants’ who were included in the study. In section 2 the nature and extent of cognitive impairment, mood disorder and physical health is analysed. Differences in performance between the exposed and control cohorts on psychometric testing are presented. In section 3 the relationships between cognitive impairment and indices of exposure are described.

SECTION 1

Recruitment rates for exposed cohort (farmers)

The total number of sheep farmers identified who had retired, reduced their workload or changed their occupation was 222 and the total number of sheep farmers identified who were still fit enough to be in employment was 212. Similar numbers of retired and working farmers were identified by written correspondence using contact details originating from British Telecom and the National Farmers Union. The majority of working farmers were identified by telephone cold calling utilising contact details held by the Wool Marketing Board; whilst just over half the sample of retired individuals (n=127) were identified via some sort of advertising support groups or word of mouth (see appendix 3 for information about study participants and their sources)

Examined Participants: Exposed cohort

Sixty seven percent (290) of retired and working farmers identified had to be excluded because they did not meet one or more of the inclusion criteria for the study, they declined to take part in the study or they had a medical condition which might have an adverse effect on cognitive function. This left a final sample of 65 retired and 79 working farmers (see appendix 4 for reasons for exclusion)

Recruitment rates for control cohort (Police)

The total number of police workers identified who had retired, reduced their workload or changed their occupation was 170 and the total number of police workers identified who were still fit enough to be in employment was 82. The majority of employed police workers were recruited via advertising whilst the majority of retired police workers were recruited via NARPO (see appendix 3).

Examined Participants: Control Group (Police)

Sixty three percent (158) of retired and working police had to be excluded because they did not meet one or more of the inclusion criteria for the study, they declined to take part in the study or they had a medical condition which might have an adverse effect on cognitive function (see appendix). A further 12 police were eligible, but not examined due to time constraints. This left a final sample of 42 retired and 40 working police who were included in this study (see appendix 4).

Demographic Information & Cohort Matching

Before analysis took place, the groups had to be matched on a number of demographic factors thought to influence cognitive performance in order to eliminate any potential confounds: age, years in education and premorbid intelligence. In order to best match the groups for age and education, some further participants (11 working farmers, 1 retired farmer, 2 working police and 1 retired policeman)

11

had to be removed from the study1. The final analysis included 211 participants, which comprised 132 exposed individuals and 79 controls. Table 1 shows the demographic information for these groups.

Table 1. Demographic information for the control and exposed participants in both the working and retired groups

 

Age Years Education

MeanStd.

Deviation MeanStd.

Deviation

exposed group

working (n=68) 51.49 8.44 11.90 2.22retired (n=64) 57.70 9.64 11.17 1.79total 54.50 9.53 11.55 2.05

control group

working (n=38) 48.39 5.32 11.84 1.39retired (n=41) 54.95 7.59 12.05 1.75total 51.80 7.34 11.95 1.58

While the groups were successfully matched for education, this was not entirely the case for age. Due to the fact that people who are retired are generally older than those who are still in employment (as is the nature of retirement), it was not possible to match the retired contingent to the working group in terms of age. However, within both the working and retired groups, the controls were matched (see table 1). As the majority of the tests used within the test battery were age-scaled (i.e. the effects of age were removed), the fact that the groups overall were not matched is not a problem. However, to alleviate any possible confounds of age, when examining non-age-scaled tests, the working and retired groups will be dealt with separately.

It was further attempted to match the groups for a measure of intelligence that may represent premorbid intelligence. The issue of estimating premorbid intelligence of a potentially cognitively impaired group is not easy. The most commonly used measures tend to include verbal measures (e.g. NART, vocab subset of the WAIS), however previous research by Mackenzie Ross et al suggested organophosphate exposed individuals have impaired verbal abilities27 and therefore these measures may not be reliable. As such, premorbid intelligence was estimated using a measure that is unlikely to have been affected by cognitive damage: matrix reasoning28. An Independent Univariate Analysis of Variance (ANOVA) showed that the controls and exposed participants were successfully matched on this measure (F<1).

1 The main reason for this was with regard to age and due to the fact that police usually retire at 55 years of age (and thus the control cohort contained few working participants over this age), whereas farmers tend to work all of their lives (resulting in many workers in this group between 55 and 70 years old).

12

SECTION 2: Exposed versus Unexposed Cohort Comparisons; The Nature and Extent of Cognitive Impairment, Mood Disorder and Physical Ill Health Following Exposure to OPs

Cognitive Tests

In order to test the possible effects of organophosphate exposure on cognitive functioning and mood, an extensive test battery was carried out utilizing tests which are known to be reliable, sensitive and are routinely used in clinical practice in the UK. This battery included psychometric tests of response speed, executive function, working memory, verbal ability, visual and auditory memory and fine motor control. The performance of the exposed and control cohorts was analyzed for each of these areas and the descriptive statistics can be found in appendix 6.

Before analysis took place the distribution of all variables was checked to determine whether normality could be assumed. It was found that some variables exhibited unacceptable skewness and/or kurtosis for parametric statistical analyses. While it was possible to correct some of these distributions by transforming the data, transformations were not successful for all measures (see appendix 5).

To minimize any potential Type I errors (finding a difference when there should not be one), Multivariate Tests of Analysis (MANOVA) were used wherever possible. When Univariate or Multivariate Analysis of Variance was used, the variables of interest were Exposure Group (which had two levels: exposed and control) and Working Status (which also had two levels: farmers and controls). When MANOVA was not appropriate2 differences in mean scores were analysed using unrelated t -tests or the non-parametric equivalent (Mann Whitney U). Unless otherwise stated, all statistical tests were two-tailed. Where a predicted difference in performance was tested using more than one psychometric test, Larzelere and Mulaik tests were applied to control for Type I errors (Howell, 1992). This involved deriving revised alpha levels by using the formula depicted below:

α revised = α original / (k-i+1)

where α original =.05; k= number of psychometric tests used; i= the rank order/strength of the original p value (rank 1 is assigned to the most significant findings).

Working Memory

Differences in working memory were analysed using Digit Span, Digit Span Backwards, Letter-Number Sequencing and Arithmetic subtests. A two way independent MANOVA revealed a significant main effect of Exposure Group on working memory (V=.20, F(4,195)=11.98, p<.001, ηp

2=.20), but no effect of Working Status and no significant interaction between these variables (F<1 for both). These results suggest that the exposed cohort were significantly impaired on measures of working memory and this pattern was similar for both the working and retired groups.

Response Speed

Differences in response speed were analysed using Digit Symbol Substitution, CALCAP (simple) and Trails A tests. The 3 variables were analysed separately, as different statistical tests were appropriate for each one.

An independent two-way ANOVA revealed that there was a significant main effect of Group on Digit Symbol scores (F(1,203)=37.10, p<.001, ηp

2=.16), but no effect of Working Status and no significant interaction between these variables (F<1 for both). This suggests that the exposed cohort were significantly impaired on response speed and this pattern was similar for both the working and retired groups.

2 i.e. when the initial assumptions of the test were not met or when non-aged-scaled scores were being analysed

13

A similar deficit was found for Trails A performance, with the exposed participants performing significantly worse than the controls in both the working and retired cohorts (U=883.50, p<.01; U=542.50, p<.001 respectively).

In addition to these tests the CALCAP (simple) results were analyzed for both the working and retired cohorts. For this test participants were scored nominally by categorizing their performance as either normal or abnormal (when performance was more than 1.5 standard deviations away from the mean). 11.1% of the exposed working participants performed abnormally on this test, compared to 3% of the working controls, however this difference was not found to be significant (χ²(1)=1.85, ns). For the retired cohort, 19.4% of the exposed participants performed abnormally, compared to only 3% of the controls. A chi-square test revealed a significant difference between the two groups’ performance (χ²(1)= 4.86, p<.05), with the exposure group performing significantly worse than the controls.

To summarise, the exposed cohort appear to be impaired on measures of response speed.

Visual Memory

Differences in visual memory were analysed using Visual Immediate Memory and Visual Delayed Memory. A two way independent MANOVA revealed the difference between the exposed and control participants to be a significant one (V=.07, F(2,203)=7.43, p=.001, ηp

2=.07). No effect of Working Status was found (V=.01, F(2,203)=1.45, ns, ηp

2=.01) and there was no significant interaction between these variables (F<1). To summarise, these results suggest that the exposed cohort were significantly impaired on measures of visual memory and this pattern was similar for both the working and retired groups.

Auditory Memory and Information Processing

Differences in auditory memory and information processing were analysed using Auditory Immediate Memory, Auditory Delayed Memory and Auditory Recognition Delayed Index Scores. A two way independent MANOVA revealed a significant multivariate effect of Exposure Group on auditory memory and information processing scores (V=.08, F(3,203)=5.49, p=.001, ηp

2=.08) which the exposed cohort performed worse than the controls. No effect of Working Status was found and there was no significant interaction between these variables (largest V=.02; largest F=1.53).

Verbal Ability

Participants’ performance on verbal ability was assessed using Vocab, Graded Naming and Comprehension. A two way independent MANOVA revealed no significant main effects or interactions (largest V=.03, largest F=1.92). These results suggest that verbal abilities are largely intact.

Executive Function

Participants’ performance on measures of executive function were examined in three ways: 1.) tests of mental flexibility and participant’s ability to switch between tasks by successfully inhibiting responses to certain stimuli were measured using CALCAP-choice, Trails B and Stroop. 2.) Participants’ ability to use successful strategies to complete tasks was examined using Verbal Fluency. 3.) Participants’ Verbal and Visual Reasoning Skills were examined using Picture Arrangement, Comprehension and Similarities.

Mental Flexibility & InhibitionMann Whitney U tests revealed that the exposed participants were significantly impaired on Trails B compared to matched controls in both the working (U=790.00, p=.001) and retired (U=696.00, p<.001) cohorts.

14

On CALCAP (choice) 23.8% of the exposed working participants exhibited abnormal performance, compared to 12.1% of the working controls, however this difference was not significant (χ²(1)=1.86, ns). For the retired contingent, 48.4% of the exposed participants performed abnormally, compared to only 12.1% of the controls. In this case a chi-square test revealed this difference to be significant (χ²(1)= 12.33, p=.001).

In addition to this the Stroop test was examined. The Stroop test is a measure of mental flexibility, in particular, the ability to switch between competing response modes. Within the working group 22.7% of the exposed participants failed this test, performing below the cut off for abnormality, compared to 2.6% percent of the control group. It was therefore easier to nominally classify these scores as either “pass” or “fail” for analysis. A chi-square test revealed a significant difference between the two groups’ performance (χ² (1)= 7.48, p<.01), with the exposure group performing significantly worse than the controls. Within the retired group 22.6% of the exposed participants preformed abnormally on the Stroop, compared to 0% of the controls. Again, a chi-square test revealed this difference to be significantly different (χ²(1) = 10.71, p=.001).

To summarise, these results suggest that the exposed cohort were significantly impaired on measures of mental flexibility and that this pattern was similar for both the working and retired groups.

Strategy Making

Mann Whitney U tests were used to investigate whether there were significant differences in mean verbal fluency scores for the exposed and control groups. The results showed that for both the working and retired groups, the controls performed significantly better than the exposed group (U=693.00, p<.001; U=526.50, p<.001 respectively) indicating a deficit in strategy making ability in the exposed cohort.

Verbal & Visual Reasoning

In order to investigate whether there was a significant effect of Exposure Group or Working Status on Verbal and Visual Reasoning Skills, scores on Similarities, Comprehension and Picture Arrangement) were entered into a two way independent multivariate analysis of variance (MANOVA). Pillai’s trace revealed no significant main effects or interactions (highest V =.04; highest F=2.52).

Visuo-Spatial Abilities

Differences in visuo-spatial abilities were analysed using Block Design and Spatial Span. A two way independent MANOVA revealed no significant main effects or interactions (largest V=.02, largest F=1.91). These results suggest that visuo-spatial abilities are largely intact.

Fine Motor Control

As organophosphates can damage the peripheral nerves, a measure of fine motor control was included into the test battery in the form of the grooved peg board. Mann Whitney U tests were used to investigate whether there was a significant deficit in the exposed cohort.

For the working cohort, the exposed participants were found to perform significantly worse than the controls on the Grooved Peg Board for both the dominant (U=489.00, p<.001) and non-dominant hand (U=623.00, p<.001).

The same pattern was found for the retired cohort (dominant hand: U=633.50, p<.001; non-dominant hand: U=697.50, p<.01).

15

Validity of Neuropsychological Test Results – Effort testing.All study participants were required to complete a brief computerised verbal memory screening test which measures a person’s effort on testing. The measure used is insensitive to all but the most extreme forms of cognitive impairment whilst being very sensitive to poor effort and exaggeration of cognitive difficulties. Low scores raise the possibility that a participants’ memory test scores does not accurately reflect their true level of ability21.

Eight study participants obtained low scores on this measure (6 retired farmers, 1 working farmer and 1 retired control subject). Two of the retired farmers had already been excluded from the analyses for other reasons. To ensure that the pattern of findings from this study were not due to poor or inconsistent effort, the remaining six subjects were removed and the data re-analysed after excluding these participants. This did not change the pattern of results.

Summary

Table 2 summarises the areas of cognitive deficit in the exposed cohort compared to the controls based on the results above.

Table 2. Areas of cognitive deficit in the exposed cohort Working Retired

Response Speed x xWorking Memory x xVisual Memory x xAuditory Memory x xVerbal AbilitiesMental Flexibility x xStrategy Making x xVerbal & Visual Reasoning AbilityVisio-Spatial AbilitiesFine Motor Control x x

Further Group Analyses

Further analyses were undertaken to address specific issues which could affect or limit interpretation of the data and any conclusions that can be drawn from this study (please see appendix 7):

(1) Could the pattern of deficit observed above have been driven by inclusion of study participants with undiagnosed acute exposure?

(2) Could the pattern of deficit observed in the initial analyses be due to selection of an inappropriate control group (ie rural policeworkers) who differ from farmers in some important way other than exposure history?

16

Could the pattern of deficit observed above have been driven by inclusion of study participants with undiagnosed acute exposure? Re-analysis of the data after removal of participants who report a history of ‘dippers flu’.

The study aim was to determine whether low level exposure to OPs causes neuropsychological and psychiatric disease. COT defined low dose exposures as those which do not produce frank toxic effects accompanied by recognised clinical symptoms of acute toxicity which occur immediately following exposure. Although it has been proposed that dippers flu may be a manifestation of acute toxicity, COT concluded that this is unproven and do not consider it an indicator of acute toxicity.

However, a number of participants (34.8% of the working cohort and 50.8% of the retired group) reported that throughout their working life they suffered repeated episodes of flu-like symptoms following exposure to OPs. The farming community refers to this phenomenon as ‘dippers flu’. The cause and nature of ‘dippers flu’ has not been established scientifically, but the symptoms have much in common with those associated with mild exposure to organophosphate compounds and appear to share a temporal relationship with exposure to sheep dip. Hence, ‘dippers flu’ may reflect undiagnosed, untreated acute toxicity.

To examine whether the patterns of deficit observed in the initial analysis could have been driven by participants with undiagnosed acute exposure, the data was re-analysed after removing all farmers who had experienced ‘dippers flu’. The same areas of deficit remained after their removal, with only one change on one measure of response speed. In the initial analysis a significantly higher proportion of farmers were observed to have abnormal CALCAP (simple) scores compared to the controls in the retired group, but not in the working group. When those with dippers flu were excluded from the analysis this pattern was reversed.

Could the pattern of deficit observed in the initial analyses be due to selection of an inappropriate control group? Exposed Cohort versus Normative Comparison Standards: Re-analysis of the data using an alternative comparison group (please see attached report).

A potential weakness of this study design which could limit the conclusions that can be drawn from the above analyses was the recruitment of rural police workers as an unexposed control group. Although matched to the farmers as far as possible in terms of characteristics which may affect cognitive function (i.e. age, gender, education level, premorbid IQ), police workers differ from farmers in terms of the exact nature of the work they undertake, lifestyle and life experiences. Differences in performance on neuropsychological testing between exposed farmers and unexposed rural police workers could be due to an unidentified confounder that was not controlled for in this study and may not reflect exposure history. Therefore, the above analyses were repeated using normative comparison standards (the neuropsychological test battery consisted of well known, reliable and clinically sensitive measures for which population test norms are available).

Please see appendix 7 for details of this re-analysis. In summary the overall findings of this study are the same whether exposed farmers are compared with rural police workers or with published test norms derived from a cross section of healthy adults in the general population. Farmers show deficits on tests of working and general memory, response speed and mental flexibility, but preserved verbal, visuo-spatial, reasoning and general intellectual functioning.

Mood Disorder

Table 3 shows the mean mood scores for the different groups. Over 40% of farmers were found to be suffering from clinically significant levels of depression and anxiety (according to the Hospital

17

Anxiety and Depression Scale) compared to less than 23% of controls. Retired farmers reporting the highest rates of distress.

Table 3 Mean scores on the Hospital Anxiety and Depression Scale for the controls and exposed participants in both the working and retired groups

  Working Retired

Mean SD Range Mean SD Range

HADS Depression Score            

Exposed Group 4.66 3.76 0-16 9.37 4.19 2-17

Control Group 2.41 2.33 0-10 3.25 2.72 0-14

HADS Anxiety Score            

Exposed Group 6.17 4.12 0-15 9 4.16 0-17

Control Group 3.76 3.09 0-8 4.96 3.39 0-9

Two-way Univariate ANOVAs showed that there were significant main effects of Exposure Group (F(1, 200)=68.42, p<.001, ηp

2= .26), Working Status F(1, 200)=30.08, p<.001, ηp2= .13) and there was

a significant interaction between these variables F(1, 200)=14.56, p<.001, ηp2= .07) for the depression

scores. For anxiety scores there were also significant main effects of Exposure Group (F(1, 200)=33.82, p<.001, ηp

2= .15) and Working Status (F(1, 200)=13.38, p<.001, ηp2= .06), however the

interaction term was not significant (F(1, 200)=2.12, ns, ηp2= .01).

Re-analysis of the data controlling for the effects of mood

As previous research has shown that depression and anxiety may be related to poor performance on certain psychometric tests30 and as there are differences in mood scores between the different exposure and employment groups, it is therefore standard practice to take account of participants’ mood scores when analysing their performance on neuropsychological tests. In order to do this the above analyses on cognitive performance were re-run with the effects of mood partialled out3. Results showed that even when the effects of depression and anxiety were removed, the pattern of impairment remained the same.

Physical Symptoms

In addition to Psychological Symptoms, measures of physical health were also taken using an extensive physical health questionnaire.

3 This was done by re-running the above ANOVAs and MANOVAs with age scaled variables, but by including depression and anxiety scores as covariates.

18

Table 4 Physical symptoms reported by study participants according to occupational group and working status.  Working Retired

  Mean Score

Standard Deviation

Mean Score

Standard Deviation

Overall Health Rating        Exposed Group 2.52 0.77 3.90 1.01Control Group 2.05 0.80 3.05 1.13

Number of Moderate-Severe Symptoms

Exposed Group 3.10 3.57 11.48 7.31Control Group 1.55 1.57 2.51 2.60

Mann Whitney U tests were used to investigate whether there were any significant differences in the mean overall health rating scores or the amount of moderate to severe physical symptoms the exposed and control groups reported having.

In both the working and retired cohorts, the exposed participants reported significantly worse general health than the controls (U=848.50, p<.01; U=653.00, p<.001 respectively). A similar pattern was found in terms of the number of moderate to severe symptoms that were reported, with the exposed group reporting more symptoms than the controls in both the working (U=990.00, p<.05) and retired (U=292.50, p<.001) cohorts. Please see appendix 8 for information about the most commonly reported symptoms in each group.

19

SECTION 3: Exposure History; Descriptive Information and the Relationship Between Cognitive Function and Indices of Exposure.

Exposure history

Participants in this study had been exposed to OPs over a number of years as a result of sheep dipping and were asked to report retrospectively on their exposure history. Unsurprisingly, some individuals had difficulty giving precise details about the duration and frequency of exposure (estimates were given instead) or the names of the chemical products used. Participants used a variety of OP products of differing compositions, some individuals using more than one product during their lifetime. Exposure history varied considerably despite participants appearing to have similar jobs (see table 5).

Table 5: Quantifying ExposureMean Standard

Deviation Range

No. years spent working with OPsWorking Group 22.87 10.74 8-49 Retired Group 25.79 14.79 5-66

No. days per year spent working with OPs Working Group 3.76 6.84 0.5-30 Retired Group 2.83 3.42 1-21

Years Since Last Dip Working Group 9.59 8.76 0-37 Retired Group 11.42 7.60 0-42

Exposure IndicesLifetime Exposure Index (days)

Working Group 87.09 176.15 8-1020 Retired Group 70.26 83.88 5-504

Esk Exposure Metric4 Working Group 134.76 473.64 0.38-2906.25 Retired Group 67.72 82.82 0.33-405.00

‘Dippers Flu’

A number of participants (34.8% of the working cohort and 50.8% of the retired group) reported that throughout their working life they suffered repeated episodes of flu-like symptoms (e.g. fatigue, muscle pain, headaches and general malaise) following exposure to OPs. The farming community refers to this phenomenon as ‘dippers flu’. A chi-square test revealed no significant difference between working and retired farmers in terms of how many reported dippers flu (χ² (1)= 3.31, ns). The farming community refers to this phenomenon as ‘dippers flu’. The cause and nature of ‘dippers flu’ has not been established scientifically, but the symptoms have much in common with those associated with mild exposure to organophosphate compounds and appear to share a temporal relationship with exposure to sheep dip.

Relationship between Cognitive Tests and Exposure Indices

Spearman correlations were used to establish whether there was a relationship between cognitive function and exposure history. The OP cohort perform significantly worse than healthy controls on a range of psychometric tests. In this section analyses are undertaken to determine whether there is a linear relationship between indices of exposure and cognitive function within the exposed cohort.

4 Based on Cherrie & Robertson Exposure Metric, 1995

20

To avoid Type 1 error (finding a relationship by chance) the number of variables entered into the correlation matrix were limited. Instead of entering scores from all of the individual subtests included in the test battery, composite z-scores were calculated so that participant’s performance in the 10 cognitive domains described in the group analyses (i.e. working memory, auditory memory, visual memory, response speed, strategy making, mental flexibility, fine motor control, verbal ability, verbal/visual reasoning, visuo-spatial ability), could be examined in relation to exposure history. The composite z-scores were calculated by turning participants’ scores on each psychological test5 into z-scores and then averaging them according to cognitive domain (please see Appendix 6 for details of the cognitive domains and associated neuropsychological tests). To avoid bias all 10 cognitive domains were entered into the analyses and not just the domains shown to be impaired in the group analyses. This enabled us to examine the false discovery rate. Instead of entering data from multiple exposure indices, duration of exposure was considered the most relevant variable. As we hypothesise that increased contact to OPs will result in worse cognitive performance, 1-tailed tests are reported.

Spearman’s correlations revealed significant, negative correlations between duration of exposure and Auditory Memory (rs = -.15, p<.05), Visual Memory (rs = -.25, p<.01), Verbal Ability (rs = -.21, p<.01) and Strategy Making (rs = -.18, p<.05) indicating an association between prolonged exposure and impairments in these areas. A significant positive correlation was also found with Fine Motor Control (rs = .20, p<.05), which as a response time measure, suggests a relationship between prolonged OP exposure and peripheral nerve damage. No other significant correlations were found. In summary, we found 5 out of 10 significant correlations, all in the expected direction6 and all consistent with the findings from the group analyses and with our hypotheses (see Table 6).

5 This was only done for the tests which produced appropriate scale, interval data. Therefore CALCAP and stroop scores were not included.6 ? For the majority of cognitive tests, low scores are indicative of poorer performance, therefore negative correlations were expected between duration of exposure and cognitive test scores; however the following exceptions apply – high scores on tests of response speed (Trail making tests A&B, Grooved Pegboard ) are indicative of slow performance and therefore impairment. Hence, positive correlations were expected between tests of response speed and exposure history.

21

Table 6. Correlations between cognitive domains and exposure measures

Duration of Exposure

Working Memory 0.01

Response Speed 0.01

Visual Memory -0.25**

Auditory Memory -0.15*

Verbal Ability -0.21**

Mental Flexibility 0.11

Strategy Making -0.18*

Verbal and Visual Reasoning -0.14

Visuo-Spatial Skills -0.13

Fine Motor Control 0.20*

** - Correlation is significant at the 0.01 level (1-tailed).

* - Correlation is significant at the 0.05 level (1-tailed).

To explore the issue of chance findings even further, we examined the relationship between all of the individual cognitive subtest scores (not composite domain scores) and duration of exposure. Twenty three cognitive variables were entered into a matrix with duration of exposure. Out of 23 comparisons, 20 were in the expected direction; and only one unexpected correlation was found between duration of exposure and performance on a vocabulary test (see Table 7). If the observed pattern of results had occurred by chance one would expect an equal number of randomly distributed positive and negative correlations in the correlation matrix. 87% of the correlations were in the predicted direction and binomial tests showed that this pattern of results was significantly different from chance (p<.001). Thus while correlations may be weak, there does appear to be a relationship between exposure and cognitive deficits that cannot be attributed to chance alone.

22

Table 7. Correlations between individual cognitive tests and exposure measures

Duration of Exposure

Digit Span 0

Digit Span Forwards -0.11

Digit Span Backwards

Transformed 0.03

Letter Number Sequencing -0.10

Arithmetic 0.04

Digit Symbol Transformed -0.18*

Trails A 0.22**

Visual Immediate -0.21*

Visual Delayed -0.20*

Auditory Immediate -0.20*

Auditory Delayed -0.11

Auditory Recognition Delayed -0.03

Vocabulary -0.22**

Graded Naming -0.14

Comprehension -0.15

Trails B 0.17

Similarities Transformed -0.08

Picture Arrangement -0.16

Verbal Fluency -0.20*

Block Design -0.12

Spatial Span -0.09

Grooved Pegboard Dominant

Hand 0.22**

Grooved Pegboard Non-Dominant

Hand 0.18*

** - Correlation is significant at the 0.01 level (2-tailed).* - Correlation is significant at the 0.05 level (2-tailed).

Potential Susceptibility to OPs: Genetic Data

As well as self report measures of exposure, genetic data were taken from each of the participants to determine their ability to metabolise OPs. The human paraoxonase 1/arylesterase enzyme (PON1) is

23

a high-density lipoprotein (HDL) that plays an important role in the detoxification of organophosphates and thus help protect against the potentially harmful effects of OPs31-34. The participants’ PON1 phenotype (Q/Q; Q/R; R/R) and arylesterase activity levels were determined and are summarized in table 8.

Table 8 : PON1 status in Farmers according to work status

  Frequency of Phenotypes Arylesterase activity

 Q/Q Q/R R/R Mean Standard

DeviationWorking 26 22 5 152.84 34.02

Retired 20 27 6 170.73 61.03

The descriptive statistics show that there are similar numbers of participants in each phenotype group, with similar levels or arylesterase. It is also worth noting that no one included in this study had arylesterase levels less than 81.5 units/ml. This indicates that there were no poor metabolisers in the exposed cohort, which is not overly surprising given the fact that this study excluded any participants with a history of acute symptoms following exposure to OPs that required medical intervention (see appendix 9 figure 1 and appendix 10 for further details of assay conditions).

24

DiscussionThe present study compared the neuropsychological performance of 132 agricultural workers with a history of low level exposure to organophosphate pesticides (insufficient to produce acute intoxication requiring medical intervention) with 79 non-exposed controls (matched for age, gender, years in education and intellectual ability). Information was also obtained about physical and mental health. As far as we are aware, this is the first clinical study of farmers to take account of the ‘healthy worker’ effect by including a cohort of agricultural workers who have retired or changed occupation on ill health grounds. The overall aim of the study was to establish whether low level exposure to OPs is associated with disabling neuropsychological and psychiatric disease. A further aim was to establish whether individuals who have retired on ill health grounds constitute a particular subgroup of individuals who are more susceptible to the effects of OPs than others.

A range of emotional, physical and cognitive problems were identified in agricultural workers. In terms of emotional and physical health, over 40% of the exposed cohort complained of clinically significant levels of anxiety and depression compared to less than 23% of controls, the highest rates of distress being found in retired farmers. Farmers also report a range of physical symptoms which they describe as being moderate to severe, the most prominent being fatigue, memory problems, joint stiffness, sleep disturbance, irritability and feeling mentally slowed down.

In terms of cognitive function, general intellectual ability, reasoning, visuo-spatial and verbal ability were relatively well preserved, but agricultural workers obtained lower scores on tests of response speed, working, verbal and visual memory, mental flexibility and fine motor control, than non-exposed controls. These differences remained after controlling for Type 1 errors, depression and anxiety. Hence, these findings are unlikely to have occurred by chance or to be due to the confounding effects of mood disorder. Few differences were found between working and retired farmers in terms of the cognitive deficits identified

Further analyses were undertaken to determine whether these findings could have been driven by (1) inclusion of individuals with a history of undiagnosed acute toxicity or (2) selection of an inappropriate control group (i.e rural police workers). However, the same areas of deficit remained even after removal of study participants who report a history of ‘dippers flu’. Furthermore, the overall findings of this study are the same whether exposed farmers are compared to rural police workers or with published test norms derived from a cross section of several thousand adults in the general population.

Statistical analyses were carried out to look at the relationship between exposure history and cognitive function. A number of significant correlations were observed between duration of exposure and verbal and visual memory, verbal ability, strategy making and fine motor control. Although the correlations were weak, they were in the expected direction, consistent with findings from the group analyses and consistent with study hypotheses. Binomial tests suggest they are unlikely to have occurred by chance.

Regarding the question of whether there is a subgroup of individuals who are more susceptible to the effects of OPs than others; differences were found between working and retired farmers on subjective, self report measures of mental and physical health, but not on objective measures of cognitive function or in terms of PON1 status. The three PON1 phenotypes were equally distributed amongst working and retired farmers and similar levels of arylesterase activity were found in each group. No one included in the study had particularly low arylesterase levels so there were no poor metabolisers in the study. Therefore, individuals who have retired on ill health grounds do not appear to constitute a subgroup of people at increased risk of developing neurobehavioural impairment following exposure to OPs.

25

In summary, both correlation and group analyses suggest a relationship may exist between low level exposure to organophosphates and impaired neurobehavioural functioning and individuals who have retired on ill health grounds do not appear to be at increased risk of suffering cognitive impairment following exposure to OPs.

Comparisons with previous researchThe present findings were compared with those of previous research concerning the neurobehavioural sequelae of exposure to OPs. However, interpreting the results of previous studies is complicated by the fact there is no clear cut definition of ‘acute’ and ‘chronic’ exposure; many studies do not provided sufficient information about exposure history; results may not be comparable because different occupational groups have been examined with different sources and routes of exposure (e.g. pest control operators, chemical plant manufacturers, farmers, greenhouse workers etc); different time frames of analyses have been used with some studies examining test performance before and after a single season of pesticide use whilst others have been looking for changes over a lifetime; and variable psychometric test batteries have been used with differing levels of sensitivity1.

Cognitive deficitsFive out of thirteen studies identified that examined the neurobehavioural effects of long-term exposure to OPs found no evidence of cognitive impairment in workers who do not have a history of acute poisoning requiring medical intervention. Three of these studies may have obtained negative results because of methodological weaknesses such as the selection of an inappropriate control group36; use of psychometric tests which lack sensitivity38; division of study participants into groups based on an assumption about a likely relationship between peripheral and central nervous system dysfunction39. The remaining studies examined pest control operators with a very short history of exposure to OPs and their results seem to indicate that if episodes of acute poisoning can be avoided (including undiagnosed, mild toxicity), significant neurological sequelae can be prevented in the short term 40-41. However it is important to note that exposed subjects in Steenland et al’s (2002)study reported more symptoms suggestive of neurological impairment than controls, though few significant differences were found between the groups on quantitative tests. The remaining eight studies report a range of deficits in workers chronically exposed to OPs. All found evidence of reduced reaction time, psychomotor or information-processing speed, but inconsistent results were reported in relation to memory functioning, with some studies finding working, visual and verbal memory deficits 35, 42-44 whilst others do not10, 37, 45-46. The study by Stephens et al (1995) is similar to the current study in that UK farmers who had been exposed to OPs in sheep dip were examined. Stephens et al (1995) found that in comparison to controls, farmers performed more poorly on tests of sustained attention, syntactic reasoning and information-processing speed and had higher rates of emotional distress10. Unlike the findings of this study, memory functioning was found to be intact. There could be two reasons for this discrepancy:

(1) Stephens et al (1995) used a limited test battery which may lack sensitivity, whereas the current study used tests which are known to be sensitive to the effects of organic brain damage and are routinely used in clinical practice.

(2) Sample differences - the participants in each study may differ in some important way, for example, all of the participants in Stephens et al (1995) study were fit enough to be in employment, whereas many of the participants in the current study had retired on ill health grounds.

The pattern of deficits observed in this cohort has more in common with that seen following acute poisoning, despite the fact that this study excluded agricultural workers with a history of acute intoxication and the study findings remained the same after removal of participants who report a history of dipper’s flu, which may reflect undiagnosed acute toxicity. Previous studies that examined persistent effects following acute poisoning, have found evidence of attentional deficits, psychomotor slowing, greater vulnerability to psychiatric disorder, memory impairment, language deficits and

26

executive dysfunction 9, 47-49, 50-56. The pattern of results observed in the current study mirrors the findings from an earlier study carried out by Mackenzie Ross et al (2007)27 of 25 UK farmers, all of whom had a history of exposure to sheep dip and had retired on ill health grounds or reduced their working activities because of poor health which they attributed to OP exposure. All had a history of dippers flu, although none of them sought medical intervention for these symptoms. The exposed farmers performed more poorly than controls on tests of memory, verbal ability, response speed and mental flexibility and over 70% suffered clinically significant levels of mood disorder.

Although many studies of workers exposed to OPs have found evidence of memory impairment following previous episodes of acute poisoning, different rates of verbal as opposed to visual memory impairment have been found by different investigators 9, 48-52. This illustrates the importance of including psychometric tests which assess both verbal and visual modalities, yet some investigators have failed to do this 39.

Of further interest is the similarity between the findings of the present study (i.e. working memory and learning deficits) and those of animal experiments. Prendergast, Terry and Buccafusco (1997, 1998) examined the effects of low-level exposure to organophosphates on memory functioning in rats and found that chronic exposure to OPs, insufficient to elicit symptoms of cholinesterase toxicity, impaired new learning in rats but not prior learning/knowledge. This impairment persisted even after withdrawal from OP exposure. AChE activity in the frontal cortex and hippocampus was suppressed (areas known to be involved in learning and memory) and hippocampal AChE activity recovered at a much slower rate than other brain regions. They conclude that extended exposure to OPs in industrial or agricultural settings may produce selective impairment of working or short term memory, but may not significantly affect long term, reference memory 57-58.

Emotional changesOver 40% of the agricultural workers in this study were found to be suffering from significant levels of anxiety and depression and this is consistent with the findings of previous studies looking at the long-term effects of exposure to OPs 10, 36, 39, 42, 53-56, 59-60. Although anxiety and depression can have a negative impact on cognitive function 30, mood disorder does not appear to be responsible for the cognitive deficits observed in this study since differences remained between exposed and unexposed participants on cognitive testing even after the effects of mood disorder were controlled for statistically.

Physical symptomsAgricultural workers in this study reported a range of physical and psychological symptoms, such as fatigue and impaired memory. Retired farmers report more severe symptoms than retired control participants. The prevalence of these symptoms in the general population is high 64, making it difficult to demonstrate unequivocally a causal link with OPs 65.

The physical symptoms reported by agricultural workers in this study are similar to those reported by other agricultural workers and occupational groups exposed to OPs. For example, the UK Veterinary Medicines Directorate (VMD) commissioned an analysis of 646 reports made to their Suspected Adverse Reaction Surveillance Scheme (SARSS) and a report was published in 2002 66. The following symptoms were frequently reported by agricultural workers following exposure to OPs: headache, dizziness, paraesthesia, fatigue, gastro-intestinal disturbance, depression, musculo-skeletal disorders, memory problems and respiratory disorders. In 2003 Tahmaz, Soutar and Cherrie examined 63 respondents to the VMD SARSS scheme and found a high incidence of symptoms consistent with chronic fatigue syndrome in sheep farmers who use OP pesticides 67. Higher fatigue scores were associated with higher exposure to OPs. An epidemiological survey carried out in the UK of 367 sheep farmers who report ill health which they attribute to exposure to OPs found a high incidence of memory problems, headache, fatigue, aching muscles and joints, irritability, word-finding difficulties, depression, anxiety, sleep difficulties. This was even after excluding individuals with a medical history which might otherwise account for their symptoms. On average the health of those with a history of acute poisoning was worse than those without such a history 12.

27

Limitations of the Present Study Exposure history and accuracy of recalled informationUnfortunately, there is no biomarker of chronic, long term exposure to organophosphate pesticides so it was not possible to quantify levels of exposure or analyse precise dose/response relationships by objective means. Instead, exposure had to be estimated via self-report. Measuring exposure in this way may be problematic as self report may be distorted by inaccuracies of memory and response bias (e.g. a tendency to over or underestimate). Given farmers in this study were being asked to provide details of work history extending back in time by 25 years (on average) and given farmers in this study showed evidence of memory impairment, the accuracy of the exposure information they provided is open to question. This error will reduce the chance of finding significant associations.

A number of significant correlations were found between indices of exposure and cognitive function, but they were relatively weak. This could reflect (1) poor recall (2) inadequate measures of exposure (many types of contamination are impossible to quantify e.g. rolling fleeces) (3) the possibility that a direct linear dose-response relationship between health and exposure history does not exist. Level of exposure is frequently assumed to be the only biologically relevant/critical variable. This leads to the expectation that linear dose-response relationships between health and exposure history will be found. Some studies find such a relationship1, whilst others do not45. However, a large number of factors can influence the toxicity of chemicals including individual vulnerability and synergistic effects of chemical combinations 32, 68. These may mediate the relationship between level of exposure and health.

Animal studies demonstrate that chemical cocktails can be more toxic than what would be predicted from the known properties of each chemical which makes up the mixture. This may be due to particular chemicals knocking out the enzymes needed to detoxify others in the mixture68,69. Genetic differences between individuals render some people more susceptible to the toxic effects of certain chemicals than others. For example, the human paraoxonase 1/arylesterase enzyme (PON1) plays an important role in the detoxification of organophosphates and helps protect against the potentially harmful effects of OPs. In humans there is considerable individual variation in the serum activity of PON1 and this is partly genetically determined and PON1 polymorphisms have been identified in farmers who report ill health which they attribute to exposure to sheep dip31-34. However, this study found no relationship between PON1 status and cognitive impairment. This may have been because there were no poor metabolisers in the exposed cohort, which is not overly surprising given the fact that this study excluded any participants with a history of acute symptoms following exposure to OPs that required medical intervention.

Another type of non-linear mechanism that may be operating with low level exposure is where neuronal damage is subtle and cumulative over time, such that symptoms develop some time after exposure following depletion of functional reserves as a result of processes such as ageing.

Study Strengths

The main strengths of this study include (1) the fact that it was a detailed neuropsychological study. Lezak considers neuropsychological approaches to be the most sensitive means of examining the effects of toxic exposure as they reveal more regarding sub-clinical effects than internal dose indicators such as levels of toxins in blood or urine70. Indeed, many toxins are metabolised and excreted quickly in the human body and may not leave biological markers to prove exposure or allow level of exposure to be determined. Hence, neuropsychological testing is a useful diagnostic tool in the assessment of exposed persons71 (2) the ability to examine the nature and extent of neurobehavioural problems in this cohort in considerable depth, using clinically sensitive measures rather than administering brief screening tests or research tools which may lack sensitivity and/or specificity. The psychometric test battery was designed to cover a range of cognitive functions and included tests which are routinely used in clinical practice for diagnostic purposes. Participants were found to have deficits in particular areas whilst other abilities appeared intact. This is an important

28

finding as some of the discrepancies noted in previous research may be due to limited test batteries being employed which do not cover all classes of cognitive function (3) the inclusion of individuals who have retired from work on ill health grounds as previous studies have focused on individuals who are fit enough to work and may have underestimated risk (4) exclusion of individuals with a past medical and psychiatric history that could otherwise account for ill health (2) the wealth of information obtained about exposure history and exclusion of those who have a history of acute exposure (5) consideration and measurement of possible vulnerability factors such as PON1 status (6) being able to compare agricultural workers neurobehavioural functioning with two comparison groups; firstly an unexposed control group, matched to agricultural workers on important variables known to affect cognitive function; and secondly with published test norms derived from a cross section of several thousand adults in the general population.

Summary & ConclusionsA range of cognitive, emotional and physical problems were identified in agricultural workers with a history of low level exposure to organophosphate sheep dip. In terms of cognitive function, general intellectual ability, reasoning, visuo-spatial and verbal ability were relatively well preserved, but agricultural workers obtained lower scores on tests of response speed, working, verbal and visual memory, mental flexibility and fine motor control, than non-exposed controls. These differences remained after controlling for Type 1 errors, depression /anxiety; after removal of study participants who report a history of ‘dippers flu’; and whether exposed farmers are compared to rural police workers or with published test norms derived from a cross section of several thousand adults in the general population. Therefore, these findings are unlikely to have occurred by chance or to be due to confounding variables.

To our knowledge the current study is the first to take account of the ‘healthy worker’ effect and included farmers who had retired on ill health grounds. Although higher rates of emotional distress and physical symptoms were reported by retired farmers few differences were found on objective measures of cognitive function or potential vulnerability factors such as PON1 status. Individuals who have retired on ill health grounds do not appear to be at increased risk of suffering cognitive impairment following exposure to OPs.

Statistical analyses were carried out to look at the relationship between exposure history and cognitive function. A number of significant correlations were observed between duration of exposure and verbal and visual memory, verbal ability, strategy making and fine motor control. Although the correlations were weak, they were in the expected direction, consistent with findings from the group analyses and consistent with study hypotheses. Binomial tests suggest they are unlikely to have occurred by chance.

Both correlation and group analyses suggest a relationship may exist between low level exposure to organophosphates and impaired neurobehavioural functioning. The cognitive deficits identified in this cohort can not be attributed to mood disorder, malingering or poor effort on testing, a history of acute exposure, or genetic vulnerability in terms of PON1 polymorphisms. The pattern of deficits identified in this study is consistent with reports from previous studies and consistent with what would be expected given the principal action of OPs (i.e. inhibition of acetylcholinesterase) and the distribution of cholinergic cell groups in the brain.

There are a large percentage of cholinergic nerves in the hippocampal complex, thalamus and amygdala 72. Animals given toxic doses of OPs have neuropathological lesions characterized by axonal degeneration in these regions of the brain. Time course studies have found that lesions extend into brain areas that were not initially affected, for up to 1 year following exposure, as a result of delayed apopotic neuronal cell death (i.e. programmed cell death involving free radical generation and oxidative stress)73. Baze (1993) reviewed available published and unpublished technical reports on Soman (a nerve gas) induced morphological changes in primates. Lesions, characterised by

29

neuronal degeneration and necrosis were seen in frontal cortex, entorhinal cortex, amygdaloid complex, caudate nucleus, thalamus, and hippocampus74. These brain regions are associated with new learning and memory, arousal, attention, executive function, response speed and emotional regulation73 the cognitive functions found to be impaired in the current study of farm workers exposed to low levels of organophosphate pesticides.

Implications

The results of this study suggest there may be a relationship between long-term low-level exposure to organophosphates and the development of neurobehavioural problems. This has implications for working practice and policies and guidelines about the use of organophosphate chemicals on the farm should be reviewed.

Follow-up studies should be carried out to determine whether symptoms persist over time, improve or worsen. At present, there are no recommended treatment protocols for individuals who report chronic ill health following exposure to OPs, so there is a need for prospective treatment trials.

It is also important to consider the possibility that clear cut dose-response relationships that might be discernable following acute exposure may not be apparent with low level exposure. Low level exposure may produce subclinical neurological injury that accumulates over time and only becomes apparent when specialised neuropsychological or neurological tests are used to evaluate patients or when neuronal reserves are depleted by processes such as ageing, thus unmasking deficits71.

30

REFERENCES

1.COT Report: Committee on Toxicity of Chemicals in Food, Consumer Products and the Environment (Department of Health). Organophosphates. Crown Copyright, 1999.

2.Albers JW, Berent S, Garabrant, DH, Giordani B, Schweitzer SJ, Garrison RP & Richardson RJ. The effects of occupational exposure to chlorpyrifos on the neurologic examination of central nervous system function: a prospective cohort study. Journal of Occupational & Environmental Medicine 2004; 48, 367-379.

3.Steenland K, Dick RB, Howell RJ, Chrislip DW, Hines CJ, Reid TM, Lehman E, Laber P, Krieg Jr EF & Knott C. Neurologic function among termiticide applicators exposed to chlorpyrifos. Environmental Health Perspectives 2000; 108, 293-300.

4.Stephens R, Spurgeon A, Calvert IA, Beach J, Levy LS, Berry H, Harrington JM. Neuropsychological effects of long-term exposure to organophosphate in sheep dip. The Lancet 1995;345 (6): 1135-1139

5.Srivastava AK, Gupta BN, Bihar V, Mathur N, Srivastava LP, Pangety BS, Bharti RS, Kumar P. Clinical, biochemical and neurobehavioural studies in workers engaged in the manufacture of quinalphos. Food and Chemical Toxicology 2000; 38; 65-69

6.Farahat TM, Abdelrasoul GM, Amr MM, Shebl MM, Farahat FM. Neurobehavioural effects among workers occupationally exposed to organophosphorus pesticides. Occupational and Environmental Medicine 2003; 60: 279-286

7.Roldan-Tapia L, Parron T & Sanchez Santed F. Neuropsychological effects of long-term exposure to organophosphate pesticides. Neurotoxicology & Teratology 2005; 27 (2), 259-266

8. C.H. Hennekens., J.E. Buring. Epidemiology in Medicine. Edited by S.L. Mayrent. Lippincott, Williams & Wilkins, USA, 1987.

9. Rosenstock, L., Keifer, M., Daniell, W.E., McConnell, R., Claypoole, K., and the Pesticide Health Effects Study Group (1991). Chronic central nervous system effects of acute organophosphate pesticide intoxication. Lancet, 338, 223-227.

10. Stephens R, Spurgeon A, Calvert IA, Beach J, Levy LS, Berry H, Harrington JM. Neuropsychological effects of long-term exposure to organophosphate in sheep dip. The Lancet 1995;345 (6): 1135-1139

11. Cherrie, J.W and Robertson, A (1995). Biologically relevant assessment of dermal exposure. Annals of Occupational Hygiene, 39: 387-392.

12. Fletcher, T., MacLehose, R., Hurley, F., Cherrie, J., Cowie, H., Jamal, G. & Julu, P. (2005) SHAPE Report: Survey of Health Complaints Among Sheep-Dippers. Department for Environment, Food and Rural Affairs.

13. Wechsler D. Manual for the Wechsler Adult Intelligence Scale-III (UK) Psychological Corporation, USA, 1998.

14. Wechsler D. Manual for the Wechsler Memory Scale-III (UK) Psychological Corporation, USA, 1998.

15. Spreen, O and Strauss, E (1991). A Compendium of Neuropsychological Tests. Oxford University Press; New York

31

16. E. N Miller. California Computerised Assessment Package Manual (CALCAP), second edition. Eric N Miller and Norland Software, USA, 2002.

17. Trenerry MR, Crosson B, DeBoe J, Leber WR. STROOP neuropsychological screening test manual. Psychological Assessment Resources, Inc, 1988.

18. McKenna P and Warrington EK. The Graded Naming Test: Manual. NFER-Nelson, Windsor, 1983.

19. Borkowski JD, Benton AL, & Spreen O. Word fluency and brain damage. Neuropsychologia 1967; 5: 135-140.

20. Trites, R. Grooved Pegboard Test, Ottawa, Ontario, Canada: Royal Ottawa Hospital, 1977.

21. Green, P. Manual for the Medical Symptom Validity Test. Green’s Publishing Inc, Edmonton, Canada, 2004.

22. Michael B. First, M.D., Robert L. Spitzer, M.D., Miriam Gibbon, M.S.W., and Janet B. W. Williams, D.S.W. (1997). Structured Clinical Interview for DSM-IV® Axis I Disorders (SCID-I), Clinician Version, User's Guide. American Psychiatric Publishing, Inc.

23. Snaith RP and Zigmund AS. The Hospital Anxiety and Depression Scale: Manual. NFER-Nelson, 1983.

24. Beck, A.T., Steer, R.A., & Brown, G.K. (1996) Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation.

25. Beck, A. T., & Steer, R. A. (1990) Manual for the Beck Anxiety Inventory. San Antonio, TX: Psychological Corporation.

26. Holmes TH and Rahe RH (1967) The social readjustment rating scale. Journal of Psychosomatic Research, 1967; 11: 213-218.

27. S J Mackenzie Ross, J S Clark, V Harrison, K M Abraham (2007). Cognitive impairment following exposure to organophosphate pesticides: A pilot study. Journal of Occupational Health & Safety: Australia & New Zealand 23 (2): 133-142

28. Griffiths, T.D. (2003). Cognitive Neurology Teaching Cases. http://www.staff.ncl.ac.uk/t.d.griffiths/introduction.pdf

29.Howell DC. Statistical Methods for Psychology (third edition). Duxbury Press, California, 1992.

30. Baddeley, A.D., Wilson, B.A., Watts, F.N (editors). Handbook of Memory Disorders. John Wiley & Sons, England, 1995.

31. Furlong, C.E., Holland, N., Richter, R.J., Bradman, A., Ho, A., Eskenazi, B (2006). PON1 status of farmworker mothers and children as a predictor of organophosphate sensitivity. Pharmacogenetics and Genomics, 16 (3) 183-190.

32. Cherry N, Mackness M, Durrington P, Povey A, Dippnall M, Smith T & Mackness B. Paraoxonase (PON1) polymorphisms in farmers attributing ill health to sheep dip. The Lancet 2002; 359: 763-4

32

33. Mackness, B., Durrington, P., Povey, A., Thomson, S., Dippnall, M., Mackness, M., Smith, T. & Cherry, N. (2003) Paraoxonase and susceptibility to organophosphorus poisoning in farmers dipping sheep. Pharmacogenetics, 13 (2), 81-88

34. Richter R.J and Furlong, C.E (1999). Determination of paraoxonase (PON1) status requires more than genotyping. Pharmacogenetics, 9: 745-753

34b Richter RJ Jarvik GP Furlong CE (2008) Determination of paraoxonse 1 (PON1) Status without the use of Toxic Organophosphate Substrates. Circ Cardiovasc Genet 1:147-152

34c Richter RJ Jarvik GP Furlong CE (2009) Paraoxonse 1 (PON1) status and substrate hydrolysis. Toxicology and Applied Pharmacology 235:1-9.

35. Bosma, H., van Boxtel, M.P.J., Ponds, R.W.H.M., Houx, P.J. & Jolles, J. (2000). Pesticide exposure and risk of mild cognitive dysfunction. The Lancet, 356, 912-913.

36 Rodnitzky, R., Levin, H. S., Mick, D.L. (1975). Occupational Exposure to Organophosphate Pesticides: A Neurobehavioral Study. Archives of Environmental Health, 30: 98-103.

37. Korsak, R., Sato, M. (1977). Effects of Chronic Organophosphate Pesticide Exposure on the Central Nervous System. Clinical Toxicology, 11 (1): 83-95

38 Albers, J.W., Berent, S., Garabrant, D.H., Giordani, B., Schweitzer, S.J., Garrison, R.P. & Richardson, R.J. (2004) The effects of occupational exposure to chlorpyrifos on the neurologic examination of central nervous system function: a prospective cohort study. Journal of Occupational & Environmental Medicine, 48, 367-379

39. IOM Report (1999): Epidemiological study of the relationship between exposure to organophosphate pesticides and indices of chronic peripheral neuropathy, neuropsychological abnormalities in sheep farmers and dippers, Phase 1,2 and 3. Institute of Occupational Medicine, Edinburgh and Institute of Neurological Sciences, Glasgow.

40. Ames, R.G., Steenland, K., Jenkins, B., Chrislip, D. & Russo, J. (1995). Chronic neurologic sequelae to cholinesterase inhibition among agricultural pesticide applicators. Archives of Environmental Health, 50 (No6): 440-444

41. Steenland,K.,Dick, R.B, Howell, R.J., Chrislip, D.W., Hines, C.J., Reid, T.M., Lehman, E., Laber, P., Krieg, Jr., E.F. & Knott, C. (2000) Neurologic function among termiticide applicators exposed to chlorpyrifos. Environmental Health Perspectives, 108, 293-300.

42. Roldan-Tapia, L., Parron, T. & Sanchez Santed, F. (2005) Neuropsychological effects of long-term exposure to organophosphate pesticides. Neurotoxicology & Teratology, 27 (2), 259-266

43. 5.Srivastava AK, Gupta BN, Bihar V, Mathur N, Srivastava LP, Pangety BS, Bharti RS, Kumar P. Clinical, biochemical and neurobehavioural studies in workers engaged in the manufacture of quinalphos. Food and Chemical Toxicology 2000; 38; 65-69

44. Farahat, T.M., Abdelrasoul, G.M., Amr, M.M., Shebl, M.M., Farahat, F.M. et al (2003) Neurobehavioural effects among workers occupationally exposed to organophosphorus pesticides. Occupational and Environmental Medicine, 60, 279-286

44. Maizlish, N., Schenker, M., Weisskopf , C., Seiber, J., Samuels, S. A (1987). Behavioural Evaluation of Pest Control Workers With Short-Term, Low-Level Exposure to the Organophosphate Diazinon. American Journal of Industrial Medicine, 12, 153-172

46. Fiedler, N., Kipen, H., Kelly-McNeil, K. (1997). Long term use of organophosphates and neuropsychological performance. American Journal of Industrial Medicine, 32, 487-96

33

47. Metcalf, D.R and Holmes, J.H. (1969). EEG, psychological and neurological alterations in humans with organophosphorus exposure. Annals of the New York Academy of Sciences, 160, 357-365.

48. Savage, E., Keefe, T., Mounce, L., Heaton, R., Lewis, J., Burcar, P. (1988). Chronic neurological sequelae of acute organophosphate pesticide poisoning. Archives of Environmental Health, 43: 38-45

49. Kurlychek, R.T and Morrow, L.A. (1989). Neuropsychological Assessment of Greenhouse Coworkers with Chronic Pesticide Exposure: Data from Pittsburgh Occupational Exposure Battery. Paper presented at the 7th Annual Meeting of the International Neuropsychological Society, Canada, February 10th 1989.

50, Reidy, T.J., Bowler, R.M., Rauch, S.S., Pedroza, G.I. (1992). Pesticide Exposure and Neuropsychological Impairment in Migrant Farm Workers. Archives of Clinical Neuropsychology, 7, 85-95.

51. Steenland, K., Jenkins, B., Ames, G ., O’Malley, M., Chrislip, D., Russo, J.(1994). Chronic neurological sequelae to organophosphate pesticide poisoning. American Journal of Public Health, 84(5): 731-736.

52. Misra, U.K., Prasad, M. & Pandey, C.M (1994) A study of cognitive functions and event related potentials following organophosphate exposure. Electromyography and Clinical Neurophysiology, 34, 197-203

53. Amr, M. M., Halim, Z. S. & Moussa, S.S. (1997) Psychiatric disorders among Egyptian pesticide applicators and formulators. Environmental Research, 73, 193–199

54. Stallones, L. & Beseler, C. (2002a) Pesticide poisoning and depressive symptoms among farm residents. Annals of Epidemiology, 12 (6), 389-394

55. Stallones, L. & Beseler, C. (2002b) Pesticide illness, farm practices, and neurological symptoms among farm residents in Colorado. Environmental Research, 90, 89-97

56. Salvi, R.M., Lara, D.R., Ghisolfi, E.S., Portela, L.V., Dias, R.D. & Souza, D.O. (2003) Neuropsychiatric Evaluation in Subjects Chronically Exposed to Organophosphate Pesticides. Toxicological Sciences, 72, 267-271

57. Prendergast, M.A., Terry, A.V.J. & Buccafusco, J.J. (1997) Chronic low-level exposure to diisopropylflurophosphate causes protracted impairment of spatial navigation learning. Psychopharmacology, 129, 183-191

58. Prendergast, M.A., Terry, A.V.J. & Buccafusco, J.J. (1998) Effects of chronic, low-level organophosphate exposure on delayed recall, discrimination, and spatial learning in monkeys and rats. Neurotoxicology and Teratology, 20, 115-122

59. Gershon, S., Shaw, F. (1961). Psychiatric Sequelae of Chronic Exposure to Organophosphorus Insecticides. The Lancet, 1371-1374

60. Levin, H., Rodnitzky, R., Mick, D. (1976). Anxiety Associated With Exposure to Organophosphate Compounds. Archives of General Psychiatry, 33: 225-228

61. Ahmed, G.M. & Davies, D.R. (1997). Chronic organophosphate exposure: Towards the definition of a neuropsychiatric syndrome. Journal of Nutritional and Environmental Medicine, 7, 169-176.

62. Davies, D.R., Ahmed, G., Freer, T. (2000). Chronic exposure to organophosphates: background and clinical practice. Advances in Psychiatric Treatment, 6, 187-192.

34

63. Davies, D.R., Ahmed, G.M., Freer T. (1999). Chronic Organophosphate Induced Neuropsychiatric Disorder (COPIND): Results of Two Postal Questionnaire Surveys. Journal of Nutritional and Environmental Medicine, 9, 123-134.

64. Baxter, P.J., Adams, P.H., Aw, T-C., Cockcroft, A. & Harrington, J.M (editors)(2000). Hunter’s Diseases of Occupations (ninth edition). Arnold, London.

65. Spurgeon, A., Gompertz, D. & Harrington, J.M. (1997) Non-specific symptoms in response to hazard exposure in the workplace. Journal of Psychosomatic Research, 43, 43-49

66. G Dunn. Report on an Analytical Study of OP Sheep Dips. UK Veterinary Medicines Directorate, 2002.

67. Tahmaz, N., Soutar, A. & Cherrie, J.W. (2003) Chronic fatigue and organophosphate pesticides in sheep farming: a retrospective study amongst people reporting to a UK pharmacovigilance scheme. Annals of Occupational Hygiene, 47, 261-267

68. Abou-Donia M, Wilmarth KR, Abdel-Rahman A, Jensen KF, Oeheme FW & Kurt TL. Increased neurotoxicity following concurrent exposure to pyridostigmine bromide, DEET, and chlorpyrifos. Fundamental Applied Toxicology 1996; 34: 201-22

69. Karalliedde LD, Edwards P & Marrs TC. Variables influencing the toxic response to organophosphates in humans. Food and Chemical Toxicology 2003; 41: 1-13

70. Lezak, M.D (2004). Neuropsychological Assessment (fourth edition). Oxford University Press.

71. Hartman, D.E (1995). Neuropsychological Toxicology: identification and assessment of human neurotoxic syndromes. Second Edition. Plenum Press, New York.

72. Mesulam, M-M (1995) Structure and function of cholinergic pathways in the cerebral cortex, limbic system, basal ganglia and thalamus of the human brain. In: Bloom, F.E., Kupfer (editors). Psychopharmacology: The fourth generation of progress (p135-146). New York: Raven Press.

73. Abou-Donia, M. (2005) Organophosphorus ester-induced chronic neurotoxicity. Archives of Environmental Health, 58(8), 484-497

74. Baze, W.B.(1993) Soman induced morphological changes: an overview in the non-human primate. Journal of Applied Toxicology, 13, 173-177.

35

APPENDICES

APPENDIX 1 – Exposure Questionnaire

APPENDIX 2 – Physical Health Questionnaire

APPENDIX 3 – Recruitment Rates

Table 9 Number of farmers identified from different sampling methods

Table 10 Included farmers and their source

Table 10b Primary reasons for retirement given by farmers examined in the study

Table 11 Number of police workers identified from different sampling methods

Table 12 Included police workers and their source

Table 12b Primary reasons for retirement given by police workers examined in the study

APPENDIX 4 – Primary Reasons for Exclusion

Table 13 Primary reasons for exclusion: Exposed cohort

Table 14 Primary reason for exclusion: control cohort

APPENDIX 5 – Abnormally Distributed Variables

Table 15 Abnormally distributed variables

APPENDIX 6 – Descriptive Statistics

Table 16 Descriptive statistics for the controls and exposed participants in both the working and retired groups on the different cognitive tests

APPENDIX 7 – Additional Analyses

Exposed Cohort versus Normative Comparison Standards: Re-analysis of the data using an alternative

comparison group.

APPENDIX 8 – Physical Symptoms

Table 17 Percentage of people who reported each physical symptom.

APPENDIX 9 – PON1 Status

Figure 1 PON1 Status of research participants

APPENDIX 10 – PON1 Assay conditions: Letter from Professor Furlong

36

Appendix 3 – Recruitment Rates

Table 9 Number of farmers identified from different sampling methods

Written Telephone Support Groups

Advertising Word of mouth

Other Totals

Retired 25 21 27 41 10 124

Semi-retired 23 21 13 19 3 2 81

Changed occupation

1 1 2 9 3 1 17

Sub-total

Retired/CO

49 43 42 69 16 3 222

Working 57 122 4 14 13 2 212

Table 10 Examined farmers and their source

Written Telephone Support Groups

Advertising Word of mouth

Other Totals

Retired 5 1 9 12 6 0 33

Semi-retired 11 4 4 10 2 1 32

Changed occupation

0 0 1 0 0 0 1

Sub-total

Retired/CO

16 5 14 21 8 1 65

Working 27 37 0 6 8 1 79

11 working and 1 retired farmer were subsequently removed from the statistical analyses for age matching purposes.

37

Table 10b: Primary reasons for retirement given by farmers included in the analysis

Reason for Retirement Included sample

Non-specific symptoms*, no attribution made

33

Non-specific symptoms attributed to OPs 16

Other ill health grounds+ 15

Total 64

*Non-specific symptoms: Participants reported suffering a range of symptoms including chronic fatigue, headaches, memory loss, lack of concentration, chemical hypersensitivity, numbness, balance problems, aches and pains and symptoms of mood disorder. The majority reported a combination of several of these symptoms.

+Other ill health grounds: These included breathing difficulties, skeletal problems (back pain, arthritis), chest pain, prostate problems and poor circulation.

38

Table 11 Number of police workers identified from different sampling methods

NARPO Convalescent Homes

Advertising Word of mouth

Totals

Retired 123 1 6 2 132

Semi-retired 4 5 1 0 10

Changed occupation

22 0 6 0 28

Sub-total

Retired/CO

149 6 13 2 170

Working 7 2 65 8 82

Table 12 Examined police workers and their source

NARPO Convalescent Homes

Advertising Word of mouth

Totals

Retired 27 0 4 0 31

Semi-retired 1 1 0 0 2

Changed occupation

4 0 5 0 9

Sub-total

Retired/CO

32 1 9 0 42

Working 6 1 29 4 40

1 working and 2 retired police workers were subsequently removed from the statistical analyses for age matching purposes.

39

Table 12b: Primary reasons for retirement given by police workers included in the analysis

Reason for Retirement Included sample

Spinal, neck, knee, shoulder, limb injury or condition

32

Reactive depression now fully resolved 3

Other ill health grounds+ 6

Total 41

+Other ill health grounds include breathing difficulties, arthritis, Gout, benign tumour, Meniere’s disease, chronic fatigue.

40

Appendix 4 – Primary Reasons for Exclusion

Table 13 Primary reasons for exclusion: Exposed cohort

Reason for Exclusion Retired Semi-retired Changed

Occupation Working TOTAL

Acute OP poisoning 4 5 4 13

Refusal 7 5 19 31

Psychiatric 2 3 5

TBI 2 5 10 17

Neurological 15 2 1 18

CV 4 2 1 7

Epilepsy 1 1

Heart 5 6 1 2 14

Lung 2 3 3 8

Liver

Kidney 1 1 2

IDDM 1 1 1 3

Endocrine 1 1

Cancer 4 1 5

Dyslexia

Alcohol 5 1 3 9

Deceased 1 1

Age 21 11 16 48

Inadequate exposure history 12 9 5 35 61

Other 11 3 1 31 46

TOTAL 97 54 9 130 290

41

Table 14 Primary reason for exclusion: control cohort

Reason for exclusion Retired Semi retired Changed Occupation

Working TOTAL

Urban 8 2 9 19

Not RIHG 17 1 1 1 20

OP exposure 25 1 7 33

Other Chemical exposure 3 1 1 5

Refusal 4 2 1 7

Age 5 10 15

Current Psychiatric 11 1 12

Retired on psychiatric 2 1 3

TBI 1 1 2

Neurological 6 1 2 1 10

CV 0

Epilepsy 1 1

Heart 2 1 3

Lung 0

Liver 1 1

Kidney 1 1

IDDM 3 2 5

Endocrine 0

Cancer 0

Dyslexia 0

Alcohol 1 1 2

Area 4 1 1 7 13

Other 3 1 2 6

TOTAL: 97 4 16 41 158

42

Appendix 5 – Abnormally Distributed Variables

Table 15 Abnormally distributed variables

 

Square Root Transformation

Natural Logarithm Transformation

Reversed and Square Root Transformation

Abnormally Distributed

Digit Span Backwards x      Digit-Symbol Substitution x      Similarites   x    Matrix Reasoning7     x  Trails A       xTrails B       xGrooved Pegboard (Dominant Hand)       xGrooved Pegboard (Non-Dominant Hand)       xStroop     xCALCAP simple xCALCAP choice       x

7 When the groups were matched on matrix reasoning scores, this was done using the transformed scores.

43

Table 16 Descriptive statistics for the controls and exposed participants in both the working and retired groups on the different cognitive tests.

Working Retired

Test Mean score

Standard deviation

Range Mean score

Standard deviation

Range

Working Memory

Digit Span Forwardsa

Exposed Group 6.3 1.21 4-9 6.02 1.18 4-9

Control Group 7 1.14 5-9 6.87 1.3 4-9

Digit Span Backwards Transformeda

Exposed Group 2.16 0.26 1.73-2.83 2.09 0.26 1.73-2.83

Control Group 2.34 0.29 1.73-2.83 2.25 0.29 1.73-2.83

Letter-Number Sequencinga

Exposed Group 9.78 2.33 3-15 9.27 2.93 1-18

Control Group 11.63 2.48 7-18 11.78 2.41 5-18

Arithmetica

Exposed Group 11.68 2.67 5-17 11 2.85 4-17

Control Group 11.32 3.06 5-17 11.82 2.89 5-18

Response Speed

Digit Symbol Transformeda

Exposed Group 2.94 0.41 2.24-4.12 2.93 0.45 2.00-4.00

Control Group 3.31 0.33 2.65-4.00 3.25 0.35 2.65-4.00

Trails A b

Exposed Group 38.51 15.66 19-122 43.38 14.14 18-99

Control Group 31.53 10.96 15-69 33.08 14.85 17-98

Visual Memory

Visual Immediate a

44

Exposed Group 91.38 17.07 61-138 91.2 15.05 61-124

Control Group 100.03 15.56 75-130 97.05 17.68 65-130

Visual Delayed a

Exposed Group 92.85 14.59 62-132 93.78 14.94 62-140

Control Group 101.12 15.15 72-144 101.75 13.81 78-129

Auditory Memory & Information Processing

Auditory Immediate a

Exposed Group 100.62 14.49 68-134 96.5 17.25 62-138

Control Group 106.89 13.34 80-138 108.37 14.59 71-138

Auditory Delayed a

Exposed Group 101.75 14.3 67-128 97.79 15.31 71-132

Control Group 105.81 12.1 86-132 109.15 14.28 74-136

Auditory Recognition Delayed a

Exposed Group 103.73 14.31 75-135 97.94 16.72 60-125

Control Group 106.84 12.81 85-125 105.49 13.78 80-135

Verbal Ability

Vocab a

Exposed Group 10.66 2.4 7-16 9.92 2.35 3-15

Control Group 10.68 2.07 7-15 11.26 1.62 8-15

Graded Naming a

Exposed Group 12.51 1.26 9-16 12.2 1.26 9-14

Control Group 12.62 1.06 11-30 12.87 1.16 10-15

Comprehension a

Exposed Group 11.14 2.68 5-19 10.42 2.57 3-15

Control Group 11 2.24 6-17 11.79 2 8-16

Mental Flexibility & Inhibition

Trails B b

Exposed Group 85.15 37.26 40-241 105.87 48.61 43-367

45

Control Group 63 15.85 28-97 78.3 38.69 35-214

Verbal & Visual Reasoning

Similarities Transformed a

Exposed Group 1.02 0.11 1.61-2.83 0.99 0.11 1.61-2.83

Control Group 1.02 0.07 2.08-2.83 1.05 0.07 2.20-2.77

Picture Arrangement a

Exposed Group 10.01 2.74 5-17 9.78 2.63 4-17

Control Group 10.65 2.86 6-17 11.24 2.67 6-18

Strategy Making

Verbal Fluency a

Exposed Group 34.06 11.5 11-62 32.16 11.81 10-61

Control Group 42.05 6.6 30-55 46.55 12.87 19-73

Visio-Spatial Abilities

Block Design a

Exposed Group 12.05 2.79 6-17 11.75 2.9 5-17

Control Group 12.26 3.3 5-17 12.08 3.07 6-18

Spatial Span a

Exposed Group 10.21 2.69 3-17 9.86 2.54 5-16

Control Group 11.05 3.35 2-17 10.73 3.03 4-17

Fine Motor Control

Grooved Pegboard Dominant Hand b

Exposed Group 86.67 17.88 60-147 98.74 28.74 59-241

Control Group 70.78 9.17 55-93 79.71 12.83 58-111

Grooved Pegboard Non-Dominant Hand b

Exposed Group 91.43 19.22 66-163 102.36 25.07 59-170

Control Group 76.68 12.3 59-170 86.72 15.32 63.124

a – where a lower score represents a deficit in performanceb – where a higher score represents a deficit in performance

46

APPENDIX 7 – Additional Analyses

Exposed Cohort versus Normative Comparison Standards: Re-analysis of the data using an

alternative comparison group.

A potential weakness of this study design which could limit the conclusions that can be drawn from

the above analyses was the recruitment of rural police workers as an unexposed control group.

Although matched to the farmers as far as possible in terms of characteristics which may affect

cognitive function (i.e. age, gender, education level, premorbid IQ), police workers differ from

farmers in terms of the exact nature of the work they undertake, lifestyle and life experiences.

Differences in performance on neuropsychological testing between exposed farmers and unexposed

rural police workers could be due to an unidentified confounder that was not controlled for in this

study and may not reflect exposure history.

Therefore, the above analyses were repeated using normative comparison standards. The

neuropsychological test battery consisted of well known, reliable and clinically sensitive measures for

which population test norms are available. For example, the Wechsler Adult Intelligence scale and the

Wechsler Memory Scale (Wechsler scales) have been developed over many years and the current

editions are the result of extensive empirical studies in the US and UK involving a standardisation

sample of over 2000 adults aged 16-90 years. The sample was divided into 13 age groups and

stratified on key demographic variables including age, sex, years in education, race/ethnicity,

geographic region. Extensive testing of reliability and validity were undertaken, including validation

studies on clinical populations (learning disability, cortical and subcortical dementias, traumatic brain

injury, multiple sclerosis, epilepsy, alcohol abuse, schizophrenia). The Wechsler scales provide

contemporary normative information and interpretive tables allowing an individual’s performance on

these scales to be compared to national norms. Test norms were also available for all other measures

included in our battery.

To determine whether organophosphates have a negative effect on cognitive function, the pattern of

performance of both exposed farmers and unexposed rural police workers was compared to what one

would expect to see in the normal population.

Wechsler Memory Scale – III (WMS-III)

The Wechsler Memory Scale – III (WMS-III) was used to assess working, visual and auditory

memory. Discrepancies between Intelligence (IQ) and memory are sometimes used to evaluate

memory functioning. IQ score can be used as an index of probable, premorbid level of memory

ability. Discrepancy scores between the IQ estimated memory performance and actual memory

performance were calculated and indicate whether the examinees ability to learn and recall

information is commensurate with what would be expected given their intellectual functioning.

47

Farmers were more likely than rural police workers to have statistically significant differences

between their IQ and memory scores and Figure 1 shows the percentage of people performing on the

WMS-III sub-tests at an impairment level only expected in 10% of the standardization sample.

Farmers are more than twice as likely to suffer from impairments on visual, working and general

memory measures than the standardization sample and police workers. In addition, retired farmers

are more likely to be impaired on auditory memory.

Figure 1: Percentage of people performing on the WMS-III sub-tests at an impairment level only expected in 10% of the population

A series of binomial tests with .1 set as the proportion of expected impairment revealed that none of

the police workers significantly deviated from the expected 10% frequency. In contrast, binomial

tests revealed that significantly more working farmers were impaired than one would have expected

on measures of visual, immediate and general memory (VI, VDIM, GM; p<.001 for all) and working

memory (WM; p<.01). Retired farmers were shown to be significantly impaired on all measures

(WM and GM p<.01, all other measures p<.001).

Wechsler Adult Intelligence Scale-III (WAIS-III)

The WAIS III was administered to assess participants’ current intellectual functioning. The

test is comprised of 14 subtests which can also be used to measure a range of specific cognitive

functions such as working memory, response speed, verbal ability, verbal and visual reasoning and

visuo-spatial ability. Figure 2 depicts the pattern of performance of study participants on the different

WAIS-III subtests and Figure 3 shows the percentage of people in each group with significant

impairment on WAIS sub-tests according to published test norms.

48

0

10

20

30

40

50

AI VI IM AD VD ARD GM WM

WMS subtest

% p

eopl

e

Farmer WorkingFarmer RetiredControl WorkingControl Retired

Expected

frequency

Double

expected

frequency

Figure 2: Performance profiles on WAIS-III sub-tests (error bars represent ±2 S.E.)

49

0

2

4

6

8

10

12

14

16

WAIS-III subtests

Ave

rage

Sco

re

Farmer WorkingFarmer RetiredControl WorkingControl Retired

Figure 3: Percentage of people significantly underperforming on the WAIS-III sub-tests

From looking at Figure 3 it appears that overall farmers performed similarly to controls on most

measures, however they were more likely to have significant impairments on Digit Span (working

memory) and Digit Symbol (response speed) than the police workers. This was true for both working

and retired farmers. These findings were further investigated in terms of what one would expect to

see in the standardization sample. Again, looking at impairment levels so severe one would only

expect to see them in 10% of the standardization sample, a series of binomial tests with .1 set as the

proportion of expected impairment were carried out. Results revealed that none of the police workers

significantly deviated from the expected 10% frequency and the only significant deviation in the

exposed cohort was on digit symbol. That is a significantly higher proportion of both working

(p<.01) and retired (p<.001) farmers were found to have severe impairments on this measure of

response speed.

Additional Tests

In addition to the abovementioned tests, further measures of response speed and executive function

were collected (Trails A & B, verbal fluency, Stroop). Figure 4 shows the proportion of people in this

study who performed with levels of impairment one would only expect to see in 20% of the general

population. A cut-off of 20% was selected because the norms for one of the tests (Trail Making

50

0

5

10

15

20

25

30

35

40

45

Vocab

Sim Arith

Dspan Inf

oCom

pLN

S PCDsy

m BD MR PA

WAIS subtests

% s

igni

fican

t WA

IS d

efic

it

Farmer Working

Farmer Retired

Control Working

Control Retired

A&B) indicated if individuals performed below the 20th percentile but did not provide any further

information.

Figure 4: Percentage of people performing at an impairment level only expected in 20% of the population

A series of binomial tests with .2 set as the proportion of expected impairment were carried out

revealed that none of the controls significantly deviated from the expected 20% frequency. In

contrast both the working and retired farmers were significantly overrepresented at this level of

impairment on all four measures (highest p=.02). This indicates that while the control group

performed in line with the general population, the OP exposed cohort showed significant impairments

on measures of response speed, mental flexibility and strategy making.

Summary & conclusions

In summary the findings above demonstrate that while police workers generally performed in line

with standardisation samples, the exposed farmers showed significant deficits in some cognitive

domains. Table 1 summarises the areas of cognitive impairment in the exposed cohort compared to

the controls. The overall findings of this study are the same whether exposed farmers are compared

with rural police workers or with published test norms derived from a cross section of healthy adults

in the general population. Furthermore, the findings are consistent with the study hypotheses and

show deficits on tests of working and general memory, response speed and mental flexibility, but

preserved verbal, visuo-spatial, reasoning and general intellectual functioning.

51

0

10

20

30

40

50

60

Trails A Trails B VerbalFluency

Stroop

Cognitive Test

% p

eopl

e

Farmer WorkingFarmer RetiredControl WorkingControl Retired

Expected

frequency

Table 1. Areas of cognitive deficit in the exposed cohort

Working

Farmers

Retired

Farmers

General Intellectual Ability

Response Speed x x

Working Memorya x x

Visual Memory x x

Auditory Memory x

Verbal Abilities

Mental Flexibility & Inhibition x x

Strategy Making x x

Verbal & Visual Reasoning Ability

Visio-Spatial Abilities

x = cognitive deficit present, a as indexed by the WM measure of the WMS – results are inconclusive on other measures.

52

APPENDIX 8 – Physical Symptoms

Table 17 Percentage of people who reported each physical symptom.

  Farmers Controls  Working Retired Working RetiredTemperature 8.70 50.00 2.56 2.50Headaches 15.94 42.19 17.95 22.50Toothaches 4.35 4.69 2.56 0.00Sensation Loss in Digits 5.80 31.25 2.56 10.00Blurred Vision 1.45 14.06 5.13 2.50Numbness 8.70 37.50 2.56 22.50Fatigue 20.29 76.56 7.69 17.50Loss of Balance 2.90 34.38 0.00 2.50Dizziness 1.45 23.44 2.56 0.00Sexual Problems 2.90 32.81 0.00 0.00Ringing in Ears 5.80 14.06 2.56 10.00Skin Complaints 7.25 26.56 2.56 5.00Urinary Problems 4.35 17.19 0.00 5.00Temporary Deafness 4.35 18.75 0.00 2.50Joint Stiffness 37.68 57.81 35.90 57.50Sleep Problems 23.19 57.81 10.26 20.00Muscle Tremors 2.90 23.44 0.00 5.00Nausea 1.45 10.94 5.13 0.00Swollen Glands 1.45 6.25 0.00 0.00Pain in Hands/Feet 13.04 25.00 2.56 7.50Memory Problems 11.59 73.44 2.56 5.00Dry Mouth 7.25 18.75 2.56 0.00Chest Pain 4.35 29.69 0.00 0.00Disorientation 1.45 15.63 0.00 0.00Muscle Weakness 4.35 43.75 0.00 5.00Muscle Twitching 1.45 17.19 0.00 2.50Burning Sensation 2.90 12.50 0.00 2.50Shortnress of Breath 5.80 29.69 0.00 0.00Decreased Sensitivity 1.45 21.88 0.00 2.50Mentally Slowed 7.25 56.25 2.56 0.00Feverish 2.90 7.81 0.00 5.00Weight Changes 1.45 17.19 2.56 10.00Gastro-Intestinal Problems 10.14 28.13 2.56 7.50Diarrhorea/Constipation 5.80 12.50 0.00 0.00Allergies 11.59 15.63 17.95 7.50Irritability 24.64 43.75 7.69 0.00Coughing 7.25 17.19 7.69 2.50Cramps 10.14 35.94 2.56 10.00Chemical Intolerance 10.14 46.88 2.56 2.50

APPENDIX 9 – PON1 Status

53

Figure 1 PON1 Status of research participants

PON1 Status of research participants

0

5000

10000

15000

20000

25000

30000

0 500 1000 1500 2000 2500 3000

Paraoxonase (Units/liter)

Dia

zoxo

nase

(Uni

ts/li

ter)

(U

(Uni

ts/li

ter)

QQQR RR

54

Appendix 10 – Letter from Professor Clement Furlong regarding PON1 assay conditions. The letter contains his response to questions raised by peer reviewers.

U N I V E R S I T Y O F W A S H I N G T O N S C H O O L O F M E D I C I N E

Departments of Genome Sciences and Medicine

Division of Medical Genetics

February 18, 2009

Box 357720

Seattle, WA 98195-7720

206-543-1193

clem@u.washington.edu

Dr Sarah Mackenzie Ross

Director of the Neuropsychological Toxicology Unit

Research Department Clinical, Educational and Health Psychology

University College London

Gower Street

London WC1E 6BT

Dear Dr. Mackenzie Ross,

Below are responses to the comments of the reviewers of your manuscript. References related to the responses are located at the end of these responses.

55

Reviewer 1

Comment 5, second paragraph.

The relationship between levels of arylesterase and phenotype has not been clearly established.

Animal model studies with wild type inbred mice and genetically modified mice either missing the paraoxonase 1 (PON1) gene or expressing one or the other human PON1-192 alloforms [glutamine (Q) or arginine (R)] on the PON1 null background have clearly established that it is primarily the plasma level of PON1 that governs sensitivity to diazoxon exposure since the two PON1192 alloforms (Q or R) have nearly equivalent catalytic efficiencies for inactivating diazoxon (Li et al. 2000) which is present in most if not all exposures (Yuknavage et al. 1997). Since the rate of phenyl acetate hydrolysis is not affected by the PON1192 Q/R polymorphism, this rate may be used as a surrogate for plasma PON1 level (Furlong et al. 2006; Richter et al. 2008). Thus, it is not the PON1192 polymorphism that is important for establishing sensitivity to diazoxon exposure, but the level of plasma PON1.

An individual’s PON1 status was initially established by plotting rates of diazoxon hydrolysis at high NaCl vs. rates of paraoxon hydrolysis also at high salt (Richter and Furlong, 1999; Jarvik et al. 2003). This analysis clearly divides a population into three phenotypes, individuals homozygous for PON1Q192, heterozygotes for PON192 (Q/R) and individuals homozygous for PON1R192. The PON1R192 homozygotes have lower average activity than the PON1Q192 homozygotes leading to the interpretation that the PON1R192 homozygotes would be more susceptible to diazoxon exposure. The experiments by Li et al. (2000) show that injected purified human plasma PON1R192 protects at least as well as purified human PON1Q192 when injected into a PON1 null mouse. The reason for this is that while the maximal velocity of the PON1R192 is lower than that of PON1Q192, it has a higher affinity, thereby resulting in a better rate of hydrolysis at low in vivo diazoxon concentrations such as those experienced in actual exposures. Thus, contrary to what the PON1 status plots would suggest, PON1R192 homozygotes are not predicted to be more sensitive to diazoxon exposure than the PON1Q192 homozygotes. This is not true for chlorpyrifos oxon exposures, where the PON1192 genotype is also important in determining sensitivity (Li et al. 2000). In no case, however, is the plasma PON1 level not important.

A recent publication by our research group has shown an excellent correlation of rates of one substrate hydrolysis vs. others, provided that the rates are analyzed separately for each PON1192 functional genotype (Richter et al. 2008). This publication also contains a table for converting rates of hydrolysis of different substrates determined under different conditions to

56

physiological rates of diazoxon or chlorpyrifos oxon hydrolysis in vivo at any in vivo substrate concentration. Thus, the rates of phenyl acetate (or several other substrates) hydrolysis measured in vitro can be extrapolated to a diazoxon sensitivity phenotype. It is worth noting that the genetic variability of PON1 is reflected primarily in predicted sensitivity to diazoxon and that, as mentioned above, diazoxon is present in most if not all exposures (Yuknavage et al. 1997). The recent papers also describe the use of non-organophosphate substrates for determining PON1 status (Richter et al. 2008, 2009). The rates of hydrolysis of these substrates can also be translated to in vivo rates of diazoxon hydrolysis.

Reviewer 2

Summary

At the end of the summary, reviewer 2 notes that the discussion of the results of animal studies were limited to the findings of only one group. To date, our research group is the only one that has done extensive studies in a genetically modified mouse model system where the relationship between PON1 status and sensitivity to organophosphorus compounds could be addressed (Shih et al. 1998; Li et al. 2000; Cole et al. 2003; Jansen et al. 2009).

Comment 8. The reference for the original PON1 status determination has been included (Richter and Furlong, 1999). A more recent reference that compares the conditions of the original assay conditions with the diazoxonase assay run under physiological conditions showed that the data points from the two determinations are highly correlated. In addition, as noted above, conversion factors determining in vivo rates of diazoxon hydrolysis from other assays are provided in (Richter et al. 2008).

Comment 15. What is meant by a poor metabolizer is an individual with low rates of phenyl acetate hydrolysis which has been shown to be a surrogate measure of plasma PON1 levels (Furlong et al. 2006; Richter et al. 2008), the primary determinant of sensitivity to diazoxon. A number of publications show PON1 status distributions for different populations that could serve as additional “control groups” with respect to what a “normal PON1 status distribution” should look like (Davies et al. 1996; Richter et al. 1999; Jarvik et al. 2000; Jarvik et al. 2003; Richter et al. 2008)

References cited

57

Cole TB, Walter BJ, Shih DM, Tward AD, Lusis AJ, Timchalk C, Richter RJ, Costa LG, Furlong CE. 2005. Toxicity of chlorpyrifos and chlorpyrifos oxon in a transgenic mouse model of the human paraoxonase (PON1) Q192R polymorphism. Pharmacogenet and Genomics 15:589-598.

Furlong C, Holland N, Richter R, Bradman A, Ho A, and B Eskenazi. 2006. PON1 status of farmworker mothers and children as a predictor of organophosphate sensitivity. Pharmacogenetics and Genomics. 16:183-190.

Jansen KL, Cole TB, Park SS, Furlong CE, Costa LG. Paraoxonase 1 (PON1) Modulates the Toxicity of Mixed Organophosphorus Compounds. Toxicol Appl Pharmacol. http://dx.doi.org/10.1016/j.taap.2009.02.001

Jarvik GP, R Jampsa, RJ Richter, C Carlson, M Rieder, D Nickerson and CE Furlong. 2003. Novel Paraoxonase (PON1) nonsense and missense mutations predicted by functional genomic assay of PON1 status. Pharmacogenetics 13:291-295.

Li W.-F., L.G. Costa, R.J. Richter, T. Hagen, D.M. Shih, A. Tward, A.J. Lusis and C.E. Furlong. 2000. Catalytic efficiency determines the in vivo efficacy of PON1 for detoxifying organophosphates. Pharmacogenetics 10:767-780.

Shih DM, Gu L, Xia Y-R, Navab M, Li W-F, Hama S, Castellani LW, Furlong CE, Costa LG, Fogelman AM, Lusis AJ. 1998. Mice lacking serum paraoxonase are susceptible to organophosphate toxicity and atherosclerosis. Nature 394:284-287.

Richter, RJ and Furlong, CE. 1999. Determination of paraoxonase (PON1) status requires more than genotyping. Pharmacogenetics 9:745-753.

Richter RJ, Jarvik GP, Furlong CE. 2008. Determination of Paraoxonase 1 (PON1) Status without the Use of Toxic Organophosphate Substrates. Circ Cardiovasc Genet 1:147-152.

Richter RJ, Jarvik GP, Furlong CE. 2009. Paraoxonase 1 (PON1) status and substrate hydrolysis. Toxicology and Applied Pharmacology. 235:1-9.

Yuknavage, K.L., R.A. Fenske, D.A. Kalman, M. C. Keifer, C.E. Furlong. 1997. Simulated dermal contamination with capillary samples and field cholinesterase biomonitoring. J. Toxicol. and Env. Health 51:35-55.

58

59