Department of Public Health Sciences

Study of Agricultural Lung Disease (SALUD) Research Design and Methods

Sample population

The workers recruited for this study will fall into three subgroups: (1) current paraquat applicators, (2) past applicators (those who worked as an applicator in the last five years but not currently working as an applicator); and (3) workers who are non-applicators (unexposed). It is expected that the subjects will be recruited from various crop plantations with different sizes, in addition to different herbicide handling operations. We expect to include at least three different crops, such as coffee, banana and palm oil. Approximately 150-200 current applicators will be frequency matched 1:2 by age, duration of service and farm or plantation to non-applicators. It is estimated that the number of current applicators will equal 150 — 200 workers, thus resulting in a total sample size of 450 — 600 subjects. Non-applicators (unexposed workers) would be involved in types of farm work, other than spraying. They will be similar to applicators in terms of physical exertion and socio-demographic characteristics.

Current applicators and non-exposed workers will be identified by the farm/plantation they work on. Workers will be selected from lists of all male current employees at the participating farms or plantations. Every effort will be made to survey selected subjects, including return visits to the farm or transportation from home to the study site if necessary.

A list of applicators from the last five years who are no longer working as applicators will also be obtained from each farm when available, and a random sample of these will be surveyed even if they no longer work on the study farm. We are interested in former applicators because they may have developed health problems and left this type of work.

Epidemiologic data collection

Recruitment/outreach efforts. Outreach efforts will be conducted by the field coordinator and study investigators at the farms/plantations in order to enroll workers into the study. This will include group meetings with farm supervisors and workers to answer all questions about the study. Meetings will be held at a time agreed upon with farm supervisors and convenient for farm workers. Individual participants will also have any specific questions answered at the time they are surveyed and informed consent is obtained.

All data collection will be done in Costa Rica by a local field team. The team will consist of a field coordinator/interviewer, interviewer, two technicians (one to administer pulmonary function testing and one to administer oxygen desaturation testing) and a field assistant. The team will set up a study site for data collection on the farm/plantation being studied, or in a nearby convenient location (e.g. health clinic). Small groups of workers (n=5-6) will be brought to the study site for interviewing and pulmonary function and exercise oximetry testing to minimize disruption of work schedules. Data collected for this study will include an interviewer-administered questionnaire, spirometry, single breath diffusion capacity (DlCO) and arterial oxygen desaturation. The interview, pulmonary function testing and oxygen desaturation testing are estimated to take 90 minutes per participant. The data collection team will travel to the farms/plantations to interview participants and administer pulmonary function tests to increase the ease of participation and response rate.

Interview questionnaire. A questionnaire will be administered to all participants by trained interviewers. The questionnaire will obtain demographic information, medical history, lifestyle characteristics (smoking habits and alcohol consumption), and occupational exposure history. Exposure data will be collected for each agricultural job including type of occupation, duration of service, exposure mechanism, protective measures used, and if applicable, exposure to pesticides other than paraquat. The questionnaire will use standardized questionnaires (e.g. ATS-DLD) where possible.

Pulmonary Function Testing. All Pulmonary Function testing will be done using instruments that conform to American Thoracic Society (ATS) published standards. A trained respiratory health technologist will perform all spirometry and single breath diffusion capacity measurements according to American Thoracic Society criteria. Pulmonary Function testing will be completed for all subjects, including spirometry and single breath diffusion capacity measurements.

Oxygen desaturation testing. Oxygen desaturation on exercise testing will be done by pulse oximetry during a modified stage one exercise test. This involves the measurement of arterial oxygen saturation by means of oximetry during an incremental exercise test performed to exhaustion on a mechanical cycle ergometer. Measurements will be taken using fingerclip sensors that record heart rate and arterial oxygen saturation (percent SpO2). Heart rate, workloads performed by subjects, exercise durations, and pre- and post-exercise saturation will be recorded.

Both the pulmonary function and exercise oximetry technicians will be blind to exposure status and other measurements.

Exposure Assessment

The exposure assessment component includes several mechanisms for measuring paraquat exposure including biological monitoring, air monitoring and observational information. Given the potential error from exposure classification in earlier studies, a job exposure matrix similar to one used by Dalvie (1999) will be used. In addition, there will be extensive exposure assessment to support the exposure classification. The proposed exposure assessment will provide a very good estimate of chronic exposure.

Study subjects. We will recruit up to 100 current herbicide applicators for exposure characterization with biologic markers. This will include up to 30 applicators per crop, since this sample size will be needed to determine the variance in exposure among applicators. In addition, we will select 50 non-applicators. The non-applicator will be selected from the same farm from which we recruit applicators with a ratio of 2:1. A non-applicator is defined as the farm worker who does not handle herbicides or pesticides on the farm but is working on the farm when urine samples are collected from applicators.

Biological monitoring. We will collect 24-hour urine samples from 100 herbicide applicators and 50 non-applicators during the day. We will collect urine samples on the day prior to spray activity, on the working (spraying) day and on one day after spraying. Multiple-day applicators will be screened out before the sampling. Since the plantations apply paraquat several times throughout a year, we will measure their exposures when they actually apply paraquat. This "real-world" exposure situation will provide non-biased measurements for the exposure assessment.

In addition to inter-personal variance of exposure, we will determine temporal variation of the exposure. We will select a subset of 30 applicators. We will collect urine before, during and after spraying for two additional times (thus yielding a total of three measurements). The temporal variance will be incorporated into the exposure characterization. We will not measure temporal variation of exposure in non-applicators.

Air monitoring. The air monitoring has two purposes. (1) We can determine airborne paraquat concentration during application. Using the information, we may determine major routes of entry — we may conclude that inhalation is not a major route of entry. This information about route of entry is important for potential control measures. Respiratory exposure measurement can determine the relative importance of inhalation on total exposure, and provide information on the usefulness of potential control measures. (2) Dust exposure can be confounding factor for respiratory health effects. Farmworkers can be exposed to high level of dust. The dust may include various hazardous components other than herbicides. Since the dust exposure may cause pulmonary effects, the measurement can provide valuable information on potential classification error and evaluate the potential contribution of other respiratory hazards.

Respiratory exposure to mixed dust and amount of paraquat in dust samples will be measured in 30 applicators. The exposure level of dust will be measured in 15 non-applicators as well (paraquat level will not be measured for non-applicators — it will be assumed to be very low).

Respiratory paraquat exposure will be concurrently measured with urine collection. Inhalable dust concentrations will be measured using an IOM sampler. The IOM sampler measures total inhalable dust, i.e. particles that are deposited anywhere in the respiratory tract. When the flow rate is 2 L/min, the IOM Sampler (SKC Inc) effectively traps particles up to 100 µm in size. The inhalable dust will be collected by 25 mm glass fiber filter paper.

Other environmental conditions including temperature, humidity and wind, will be measured. Other working variables, such as application equipment, amount of application handled/sprayed, personal protective equipment, whether workers mix and load in addition to apply paraquat, type of clothing worn in addition to PPE, will be recorded. In addition, a simple questionnaire will be used to collect their working experience, application methods, use of PPE and general health status.

Characterize current and cumulative paraquat exposure among farm workers. Current and cumulative paraquat exposure will be characterized by extensive crop and job exposure matrix. We will administer a questionnaire to all workers to collect information about current and previous herbicide exposure. The information will include crop, job description, duration of job, intensity of application, and experience of skin burn. Work history and historical job description will be collected from farms/plantations, when available. For cumulative exposure estimation, we will collect information about all lifetime jobs. In addition, we will utilize estate records on spraying, when available.

The herbicide exposure will be determined by crop and job specific exposure matrix, as proposed by London and Myers (1997). The job exposure matrix will be modified for herbicides. Lifetime and short term exposure will be determined by exposure duration and weights for determinant factors.