Acute Illnesses Associated With Pesticide Exposure at Schools

Context Pesticides continue to be used on school property, and some schools are at risk of pesticide drift exposure from neighboring farms, which leads to pesticide exposure among students and school employees. However, information on the magnitude of illnesses and risk factors associated with these pesticide exposures is not available.

Objective To estimate the magnitude of and associated risk factors for pesticide-related illnesses at schools.

Design, Setting, and Participants Analysis of surveillance data from 1998 to 2002 of 2593 persons with acute pesticide-related illnesses associated with exposure at schools. Nationwide information on pesticide-related illnesses is routinely collected by 3 national pesticide surveillance systems: the National Institute for Occupational Safety and Health’s Sentinel Event Notification System for Occupational Risks pesticides program, the California Department of Pesticide Regulation, and the Toxic Exposure Surveillance System.

Main Outcome Measures Incidence rates and severity of acute pesticide-related illnesses.

Results Incidence rates for 1998-2002 were 7.4 cases per million children and 27.3 cases per million school employee full-time equivalents. The incidence rates among children increased significantly from 1998 to 2002. Illness of high severity was found in 3 cases (0.1%), moderate severity in 275 cases (11%), and low severity in 2315 cases (89%). Most illnesses were associated with insecticides (n = 895, 35%), disinfectants (n = 830, 32%), repellents (n = 335, 13%), or herbicides (n = 279, 11%). Among 406 cases with detailed information on the source of pesticide exposure, 281 (69%) were associated with pesticides used at schools and 125 (31%) were associated with pesticide drift exposure from farmland.

Conclusions Pesticide exposure at schools produces acute illnesses among school employees and students. To prevent pesticide-related illnesses at schools, implementation of integrated pest management programs in schools, practices to reduce pesticide drift, and adoption of pesticide spray buffer zones around schools are recommended.

Exposure to pesticides in the school environment is a health risk facing children and school employees. Despite efforts of several organizations and laws in several states to reduce pesticide use at and around schools,¹ pesticides continue to be used in schools.² Another source of pesticide exposure at schools is from pesticides used on farmland contiguous to school facilities. However, as a result of the work of the US Environmental Protection Agency (EPA), advocacy groups, universities, state regulators, the pest control industry, and others, and laws or strong voluntary programs in several states, pesticide use has been reduced in some school districts.³

Currently, there are no specific federal requirements on limiting pesticide exposures at schools. Under the Federal Insecticide, Fungicide, and Rodenticide Act, pesticides must be registered with the EPA before they are sold or distributed.⁴ The Food Quality Protection Act⁵ of 1996 amended the Federal Insecticide, Fungicide, and Rodenticide Act, bolstering the protection of children through requiring that pesticides used on foods produce no harm. However, there are no specific provisions in these laws about the use of pesticides at schools.^1,6

The Federal Insecticide, Fungicide, and Rodenticide Act is often supplemented by more stringent state pesticide laws to protect children from pesticides at schools. For example, 18 states recommend (n = 6) or require (n = 12) schools to use integrated pest management strategies and 7 states restrict pesticide applications in areas neighboring a school.⁷ However, there are still large gaps throughout the country where children may not be afforded adequate protection.^1,8

Pesticide poisoning is a commonly underdiagnosed illness in the United States today. The clinical findings of acute pesticide poisoning are rarely pathognomonic but instead can resemble acute upper respiratory tract illness, conjunctivitis, or gastrointestinal illness, among other conditions. Detailed description of the diverse syndromes associated with different types of pesticides is available.⁹

Although some information about acute illnesses associated with pesticide exposures at schools is available,^10,11 there has not been an effort to provide a nationwide summary of this health problem. To estimate the magnitude of and the risk factors for pesticide-related illnesses associated with exposures at schools, we examined information from state-based pesticide poisoning surveillance systems (the National Institute for Occupational Safety and Health’s Sentinel Event Notification System for Occupational Risks [SENSOR] pesticides program and the California Department of Pesticide Regulation [CDPR]), and the Toxic Exposure Surveillance System (TESS), which is a national database of all calls made to poison control centers and is maintained by the American Association of Poison Control Centers.^12,13

School employees, parents, and students who developed acute pesticide-related illnesses from pesticide exposure at child care centers and elementary and secondary schools from 1998 to 2002 were identified (Table 1). Data were obtained from states participating in the SENSOR pesticides program (California, Washington, Texas, Florida, Louisiana, New York, Oregon, and Michigan), CDPR (California), and TESS (all US states and District of Columbia, with the exception of Hawaii). The data used in these analyses were surveillance data and as such are exempt from consideration by the human subjects review board and need for informed consent. Integrating data from these 3 surveillance systems provides the best available understanding of the problem of pesticide poisoning at schools. The states participating in the SENSOR and CDPR programs obtain information from multiple sources (government agencies, poison control centers, and reports from health care organizations) and conduct active case follow-up.¹² In addition, all cases identified by the CDPR are referred to the relevant county agricultural commissioner who investigates the exposure circumstances.^10,12 The TESS data are provided by approximately 67 US poison control centers.¹³ Approximately 13% of their calls come from physicians treating patients who are exposed and 87% come from patients or their relatives.^12,13

Cases were included if health effects developed subsequent to pesticide exposure and if these effects were consistent with the known toxicology of the pesticide product, as determined by state surveillance professionals (SENSOR and CDPR cases) or a poison control center specialist (TESS cases). The states participating in the SENSOR pesticides program adopted a standardized case definition in 1998, and CDPR uses a similar case definition. Briefly, the case definition required information on pesticide exposure, health effects, and evidence supporting an association between the pesticide exposure and the health effects. A full description of the standardized case definition has been previously published.¹² Identification of TESS cases relied on the experience and judgment of the poison control center specialist managing the specific case. Multiple cases exposed in a single exposure incident were identified as 1 exposure event. Exclusion criteria included exposure to substances other than pesticides, suicides, intentional abuse, and malicious use.

SENSOR and CDPR primarily capture work-related pesticide poisoning cases, whereas TESS primarily captures non–work-related cases (Table 1). Detailed information on work-related cases was provided by SENSOR and CDPR only. The SENSOR and CDPR cases were further classified into exposure to pesticides applied on school grounds when indoor and outdoor pesticide applications on school grounds resulted in illness, and to pesticide drift when pesticide drift from applications to neighboring farmland resulted in illness among students and school employees.

For the present analyses, the toxicity category of the pesticide product was retrieved from a data set made available by the EPA. The EPA assigns acute toxicity category I to the most toxic pesticide products and category IV to the least toxic pesticide.¹⁴

Illness severity was categorized for SENSOR and CDPR cases using standardized criteria.¹⁵ State agencies classified severity for the cases they identified in 2001 and 2002. Two authors (W.A.A. and G.M.C.) assigned severity to 1998-2000 SENSOR cases, all CDPR cases, and all TESS cases.¹⁶ High severity includes cases in which the illness or injury is severe enough to be considered life-threatening and commonly involves hospitalization to prevent death. Signs and symptoms include seizures and pulmonary edema. Moderate severity illness or injury includes cases of less severe illness or injury often involving systemic manifestations requiring treatment. The individual is able to return to normal functioning without any residual disability. Low severity illness or injury typically resolves without treatment and is often manifested by skin, eye, or upper respiratory tract irritation.¹⁵

Data quality control procedures included the elimination of duplicates between SENSOR (California) and CDPR, and between SENSOR and CDPR combined and TESS. To detect duplicates between SENSOR and CDPR combined and TESS, a case-by-case comparison was performed when a reporting source for SENSOR and CDPR cases was a poison control center. Cases that matched each other on state, date of exposure, age, sex, and pesticide name were assumed to involve the same individual. Such individuals were included only once in the state agency totals. Six CDPR and 8 TESS duplicates were deleted.

SAS release 8.02 (SAS Institute Inc, Cary, NC) and Epi Info version 3.2.2 (Centers for Disease Control and Prevention, Atlanta, Ga) were used for data management and statistical analysis. Age was stratified into children (<18 years) and adults (≥18 years).

Illness incidence rates among children were calculated. Rate numerators were obtained by summing the number of ill children reported by year, and denominators were obtained from the US Census data¹⁷ by summing the number of children in the corresponding state and year. Denominators were adjusted by subtracting estimates of preschoolers not attending organized child care centers¹⁸ and home-schooled children.¹⁹

Illness incidence rates among school employees were calculated for SENSOR and CDPR cases only. Denominators were obtained from the Current Population Survey²⁰ by summing the number of full-time equivalents employed in schools in states and years that contributed to the numerator. Non–work-related cases (eg, parents) and cases with unknown work-related status, which included all TESS cases, were not included in these calculations.

We used odds ratios (ORs) to assess whether age, sex, acute toxicity pesticide category, surveillance system, or site of pesticide applications were associated with severity of illness. Odds ratios, 95% confidence intervals (CIs), χ² tests, and P values were calculated using the Epi Info Statcalc utility. SAS release 8.02 was used to calculate the Poisson regression test for trends in incidence rates across the years of exposure. P≤.05 was considered statistically significant.

From 1998 to 2002, 2593 individuals were identified with acute pesticide-related illnesses associated with pesticide exposures at schools. SENSOR identified 147 cases (6%), CDPR identified 259 cases (10%), and TESS identified 2187 cases (84%) (Table 2). Most illnesses reported by SENSOR (n = 96, 65%) and CDPR (n = 158, 61%) were adults, whereas most cases reported by TESS were children (n = 1831, 84%). Among the 2181 persons with known exact age, the mean age for children was 9.5 years (range, 0.5-17.2 years) and the mean age for adults was 36.1 years (range, 18-76 years).

Three cases of high severity illness were identified. There were no fatalities reported. The odds of high and moderate severity illness were higher among cases reported by SENSOR and CDPR (15%) compared with TESS (10%) (OR, 1.6; 95% CI, 1.1-2.2), among adults (18%) compared with children (8%) (OR, 2.6; 95% CI, 2.0-3.5), and among females (12%) compared with males (8%) (OR, 1.5; 95% CI, 1.2-2.0). Moderate severity illness was more common (Table 3) among those exposed to fumigants (n = 4, 40%), herbicides (n = 41, 15%), insecticides (n = 83, 9%), and disinfectants (n = 101, 12%). Table 4 describes symptoms of high and moderate severity cases.

Insecticides were associated with 895 illnesses (Table 2). The most frequent insecticides were pyrethrins (n = 119, 13% of all insecticides), chlorpyrifos (n = 116, 13%), malathion (n = 84, 9%), diazinon (n = 78, 9%), and pyrethroids (n = 47, 5%). Disinfectants were associated with 830 illnesses. The most frequent disinfectants were sodium hypochlorite (n = 175, 21% of all disinfectants), phenol compounds (n = 175, 21%), pine oil (n = 104, 13%), and quaternary ammonium compounds (n = 81, 10%). Repellents were associated with 335 illnesses, including naphthalene (n = 136, 41%) and diethyl toluamide (DEET, n = 127, 38%). Herbicides were associated with 279 illnesses, including glyphosate (n = 100, 36%), 2,4-dichlorophenoxyacetic acid (n = 53, 19%), and pendimethalin (n = 40, 14%).

Information on the toxicity category of pesticides associated with illnesses was available for 1686 cases (Table 3). Children were less likely to be exposed to toxicity category I pesticides compared with adults (14% of children and 42% of adults, P<.001). The odds of high and moderate severity illness were higher among cases exposed to toxicity category I (18%) than cases exposed to toxicity category III pesticides (12%) (OR, 1.5; 95% CI, 1.1-2.2). The pesticide active ingredients associated with high and moderate severity illness are shown in Table 5.

The overall incidence rate among children for 1998-2002 was 7.4 cases per million children (Table 6). The yearly incidence rates increased from 1998 through 2002 for preschool children (P<.001), school-aged children (P = .002), and all combined (P<.001). The overall incidence rate among adults was 27.3 cases per million full-time equivalents (Table 7), and the yearly incidence rates decreased from 1998 through 2002 (P<.001).

The SENSOR and CDPR results are combined (Table 2) because the case definition and level of detail are similar. A total of 406 persons were exposed to pesticides in 173 events for a mean of 2.3 cases per exposure event (range, 1-61 cases). Eleven exposure events accounted for 208 cases (51%). The 244 work-related cases were exposed in 155 events.

Occupational Illnesses. Among the 244 work-related cases, 144 (59%) were not applying pesticides, 93 (38%) were applying or handling pesticides, and 7 (3%) had no information available. Among the 144 employees not applying pesticides, 96 (67%) were exposed to pesticides applied on school grounds and 48 (33%) were exposed to pesticide drift from neighboring farmland. Sixty-three nonapplicator illnesses (44%) were among teachers. Among the 93 school employees who were applying or handling pesticides, there were 41 custodians and gardeners, 26 food preparation workers, 7 teachers, 7 maintenance workers, and 12 unspecified school employees.

Illnesses Associated With Exposure to Pesticides Applied on School Grounds and Pesticide Drift From Farmland. A total of 281 cases (69%) that were reported to SENSOR and CDPR were exposed to pesticide applications on school grounds (Table 8). Insecticides (n = 156, 56%) and disinfectants (n = 99, 35%) accounted for most of the cases. The most common active ingredients were diazinon (n = 64, 23%), sodium hypochlorite (n = 47, 17%), chlorpyrifos (n = 40, 14%), quaternary ammonium compound (n = 38, 14%), and malathion (n = 14, 5%).

A total of 125 cases (31%) were exposed to pesticide drift. Insecticides accounted for 114 cases (91%) and fumigants for 9 cases (7%). The most common active ingredients were chlorpyrifos (n = 28, 22%), methamidophos combined with chlorothalonil and propargite (n = 25, 20%), mancozeb combined with glyphosate (n = 20, 16%), cyfluthrin combined with dicofol (n = 16, 13%), and malathion (n = 13, 10%).

Exposure via pesticide drift compared with pesticides applied on school grounds did not increase the odds of high and moderate severity illness (OR, 0.6; 95% CI, 0.3-1.2; P = .09). A higher proportion of children compared with adults were exposed via drift from neighboring farmland (40% vs 25%, P = .001).

These findings indicate that pesticide exposures at schools continue to produce acute illnesses among school employees and students in the United States, albeit mainly of low severity and with relatively low incidence rates. Illnesses were associated with pesticides applied on school grounds and with pesticide drift from neighboring farmland. The pesticide exposures at schools might be associated in part with several factors: a lack of federal and state regulations regarding pesticide usage in schools¹; regulatory noncompliance by school management, school employees, and pesticide applicators in states in which regulations and recommendations have been passed; and insufficient involvement of stakeholders (eg, parents, teachers, students, school administrators, pest managers).⁶

We found that the pesticide poisoning incidence rates among children increased during the period of our report. Given that 40% (n = 59) of SENSOR and CDPR cases involving children were exposed to pesticide drift and, given increasing suburban sprawl, this trend among children might be related to an increased number of schools situated next to farmland.⁶ Additional studies are needed to confirm this hypothesis. Hypotheses for the decreasing trend in illness rates among school employees include changes in pesticide use practices and increased awareness of the toxic effects of pesticides.

Incidence rates among school employees were found to be higher than incidence rates among children. Possible explanations include school employees are called to protect children when incidents occur, whereas students are often quickly evacuated; school employees are at schools for more hours compared with students; and some school employees handle or apply pesticides.

Based on SENSOR and CDPR data, most cases of acute pesticide-related illnesses were associated with pesticides applied on school grounds (n = 281, 69%). Repeated pesticide applications on school grounds raise concerns about persistent low level exposures to pesticides at schools. It is known that some pesticides degrade slowly when they are not exposed to sun, rain, and bacterial action in the soil.^21-24 In addition, pesticide residues on the school grounds might be tracked into school buildings by students and school employees. The chronic long-term impacts of pesticide exposures have not been comprehensively evaluated; therefore, the potential for chronic health effects from pesticide exposures at schools should not be dismissed.²⁵ Unfortunately, the surveillance methods used in our report are inadequate for assessing chronic effects.

Although insecticides were most frequently associated with pesticide-related illnesses (n = 895, 35%), we found that exposure to disinfectants at schools might also be a cause for concern. First, disinfectants accounted for 830 (32%) of 2593 total cases and for 101 (37%) of 275 moderate severity cases. Second, 259 (56%) of 461 cases of disinfectant exposure with toxicity category available were of toxicity category I. Finally, most of the disinfectants associated with moderate illnesses were products commonly used at schools (sodium hypochlorite and quaternary ammonium compounds).

We also found acute illnesses associated with exposure to pesticide drift from neighboring farmland. These exposures might have resulted from pesticide applicators not complying with pesticide labels, regulations, and/or guidance to avoid pesticide spray drift, or lack of federal and state regulations regarding pesticide application around schools. Additionally, pesticide drift from neighboring farm fields might increase pesticide exposure inside schools. Some studies^26-29 suggest that dwellings adjacent to fields can be contaminated by pesticide drift during applications and subsequent wind recirculation of dust from the fields.

To prevent illnesses associated with pesticide applications on or near school grounds, there is a need to reduce pesticide use. This can be accomplished by implementing integrated pest management at schools and using methods that reduce pesticide drift from farmland. Integrated pest management programs can reduce pesticide use at schools.^3,30 Integrated pest management is endorsed by the EPA,³ National Parent Teacher Association,³¹ National Education Association, and other organizations. The elements of integrated pest management are detailed in the Box. Useful guidance and references on integrated pest management in schools are widely available.^3,32 Some disadvantages of integrated pest management implementation include the requirements of more involvement of school employees, parents, and students, and the need to be educated on pest biology and integrated pest management. Finally, some economic investment is usually required at the outset of an integrated pest management program. However, over the long term, the costs of integrated pest management have been found to be lower than traditional pest control.^3,30

Pesticides Applied on School Property
Implement school integrated pest management programs:

• Monitor for pest problems.

• Identify the sources of any pest problems.

• Eliminate the sources of any pest problems, using pesticides only as a last resort. Use nontoxic methods, such as ensuring sanitary conditions and structural integrity.

• If nontoxic pest control methods are impractical or unsuccessful, then use pesticides having the lowest possible toxicity. Pesticides in US Environmental Protection Agency toxicity categories I and II should be avoided if possible. If pesticides are used:

  Provide prior written notification of the application.

  Post notices in designated areas at the school.

  Students and staff should not be present during pesticide applications.

  Restrict entry into a previously treated area for a specified duration following an application.

• Call a poison control center or seek medical attention if pesticide-related illnesses arise.

• Trained and qualified workers should handle and apply pesticides. They must be provided with appropriate safety equipment.

• Put the school’s policy on pest control in writing and distribute it to school stakeholders periodically (eg, at the beginning of the school year).

• Involve and train stakeholders (school management, parents, teachers, students, and pesticide applicators).

Pesticide Drift From Neighboring Farmland
Reduce or eliminate application methods that result in drift.
Timing of pesticide applications. Applications should be performed when students and school employees are not present.
Farmers and pesticide applicators should comply with labels, regulations, and guidance to avoid pesticide spray drift.
Pesticides should be applied by trained applicators.
Establish and enforce nonspray buffer zones around schools. Size of buffer zone depends on toxicity of pesticide, type of application (ground or aerial), and weather conditions. For example, 7 states require buffer zones ranging from 300 feet to 2.5 miles around schools.

Underreporting
Improvement in pesticide poisoning surveillance is needed. Every state should implement an acute pesticide-related illness surveillance system.
Acute pesticide-related illnesses should be a reportable condition in all states.

We tried to identify illness rate differences among children across states with different integrated pest management laws (mandatory, voluntary, without laws). However, this comparison was not meaningful because these laws have tremendous variation across states in terms of coverage, enforcement, and implementation. Additionally, 40% of cases among children in SENSOR and CDPR were exposed to pesticide drift. A similar proportion of children in the entire data set might have been exposed to pesticide drift but these cases could not be identified in TESS. Integrated pest management practices in schools are not designed to prevent exposures to pesticide drift. There were too few SENSOR and CDPR cases involving onsite applications in schools (n = 281) to assess integrated pest management laws.

Our findings are subject to at least 3 limitations. First, these results should be considered low estimates of the magnitude of the problem because many cases of pesticide poisoning are likely not reported to surveillance systems or poison control centers. Individuals who do not seek medical care or report their illness to a surveillance system or a poison control center will not be identified. Even when individuals seek medical care, their illness may not be recognized as pesticide-related, because of the nonpathognomonic nature of the signs and symptoms and because clinicians receive little training on these illnesses.^33,34 Second, although all of these cases met the definition criteria, the possibility of some false-positives cannot be excluded. Given both the nonspecificity of the clinical findings of pesticide poisoning and the lack of a criterion standard diagnostic test, some illnesses temporally related to pesticide exposures may be coincidental and not caused by these exposures. Third, although the case definition was similar, some characteristics of the populations reported by these 3 systems were different. TESS was efficient in capturing data for children, but it did not collect information on occupation, work-relatedness, and the activity the person was performing when exposed to pesticides. The SENSOR and CDPR data apply to 8 states and principally identify work-related cases. Not all states participating in SENSOR collect information on nonoccupational cases; therefore, many cases among children were likely missed by SENSOR and CDPR. None of these data sources are comprehensive. The literature suggests that less than one third of poisoning cases treated in health care facilities are reported to poison control centers and in states where SENSOR and TESS systems are in place, TESS identified only 10% of the cases identified by SENSOR.³⁵

In conclusion, despite the limitations of these 3 surveillance systems, our report is useful in providing national estimates of the magnitude of pesticide-related illnesses among school employees and students, and in identifying the risk factors that should be targeted for prevention. Strategies recommended to reduce pesticide exposures at schools include adopting integrated pest management programs and using methods to reduce pesticide drift from farmland.

Corresponding Author: Walter A. Alarcon, MD, National Institute for Occupational Safety and Health, 4676 Columbia Pkwy, Mail Stop R-17, Cincinnati, OH 45226 (walarcon@cdc.gov).

Author Contributions: Dr Alarcon had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Alarcon, Calvert.

Acquisition of data: Calvert, Blondell, Mehler, Sievert, Propeck, Tibbetts, Becker, Lackovic, Soileau, Das, Beckman, Male, Thomsen, Stanbury.

Analysis and interpretation of data: Alarcon, Calvert.

Drafting of the manuscript: Alarcon, Calvert.

Critical revision of the manuscript for important intellectual content: Alarcon, Calvert, Blondell, Mehler, Sievert, Propeck, Tibbetts, Becker, Lackovic, Soileau, Das, Beckman, Male, Thomsen, Stanbury.

Statistical analysis: Alarcon.

Administrative, technical, or material support: Alarcon, Calvert, Blondell, Mehler, Sievert, Propeck, Tibbetts, Becker, Lackovic, Soileau, Beckman, Male, Thomsen, Stanbury.

Study supervision: Calvert.

Funding/Support: This study was supported by the US government through the US Environmental Protection Agency and the Centers for Disease Control and Prevention, which employs Drs Alarcon, Calvert, and Blondell.

Role of the Sponsor: The National Institute for Occupational Safety and Health/Centers for Disease Control and Prevention (CDC) designed and conducted the study; collected, managed, analyzed, and interpreted the data; and prepared, reviewed, obtained external peer review, and approved the manuscript.

Disclaimer: The reviews expressed and the results/conclusions reached within this article do not necessarily reflect the opinions of the CDC, US Environmental Protection Agency, or each author’s state agency.

Acknowledgment: We thank Ximena Vergara (Public Health Institute, Oakland, Calif) who provided support in data management in the California Department of Health Services; Marty Petersen (National Institute for Occupational Safety and Health/CDC, Division of Surveillance, Hazard Evaluation and Field Studies) who provided support in statistical analysis; Donald Baumgartner (US Environmental Protection Agency, Region V), Sherry E. Jones (National Center for Chronic Disease Prevention and Health Promotion/CDC, Division of Adolescent and School Health), and John Palassis (National Institute for Occupational Safety and Health/CDC, Education and Information Division) who provided a comprehensive review of this article; and Jia Li (National Institute for Occupational Safety and Health/CDC, Division of Surveillance, Hazard Evaluation and Field Studies) who provided information on full-time equivalents from the US Current Population Survey.