Our scoping review was performed according to the PRISMA extension for scoping reviews (PRISMA-ScR) and accompanying checklist from Tricco et al. (2018). In Supplementary Appendix 1, we provide a guide to which pages address the specified “reporting items” from the PRISMA-ScR checklist. The protocol for this review was registered with Open Science Framework (https://doi.org/10.17605/OSF.IO/H8JZQ).

Original peer-reviewed articles published in English language journals from 1 January 2000 to 22 May 2024, were obtained through a comprehensive literature search from the bibliographic databases MEDLINE, EMBASE, Agricultural & Environmental Science Collection, CAB Abstracts, and Scopus. The final search strategy for all databases can be found in Supplementary Appendix 2.

All identified articles from the searches were transferred to screening and data extraction software (Covidence Systematic Review Software, Veritas Health Innovation, Melbourne, Australia), and duplicates were removed. The search of all databases yielded 6,154 potentially relevant articles.

Eligible articles were screened for inclusion and exclusion criteria (Table 1). To be eligible for inclusion, the original peer-reviewed articles had to meet five criteria developed by authors CRC and RJN. First, studies needed to have been conducted in the United States or Canada. Mosquito control in other countries differs substantially in methodology in large part because of differing endemic mosquito species and lack of stringent regulations and guidelines concerning appropriate insecticide applications and were thus excluded (Carrasco et al. 2019, Gorris et al. 2021). Second, this review aims to classify exposure to specific insecticides used for adult mosquito control. Studies involving insecticides other than organophosphates (e.g., pyrethroids) were excluded. Articles that examined exposure to multiple insecticides that also included organophosphates were included, but only data related to organophosphate exposure were included. Third, studies that examined insecticide exposure via routes other than ground or aerial spraying (e.g., bed nets, indoor spraying, indoor coils) were excluded given these methods are not typically used for adult mosquito control in the United States or Canada. Fourth, studies needed to report a human health effect associated with exposure to select insecticides. Studies conducted in vitro and animal studies were excluded. Fifth, only peer-reviewed articles written in English were included. Nonsystematic and systematic review articles were included. Articles identified through the article screening process not included in the original literature search that also met inclusion criteria were included. Presentations, white papers, grey literature, abstracts, and trade articles were excluded.

A team of six independent reviewers screened all relevant studies from the literature search for potential inclusion using Covidence software. In the first round of screening, a team of five independent reviewers (ZT, SKL, NF, JR, and RJN) assessed titles and abstracts of all included records using the predefined inclusion and exclusion criteria (Table 1). Each record was independently screened by random combinations of two reviewers and any discrepancies in eligibility were resolved by a third reviewer (CRC). The final round of screening included full-text review for inclusion and exclusion criteria using the same protocol described for title/abstract screening.

Data were extracted for selected articles and included the following: characteristics (e.g., title, date of publication, citation information, DOI), class of insecticide (e.g., organophosphate or pyrethrins/synthetic pyrethroids), setting characteristics (e.g., distribution methods, location, temperature, and weather conditions), methodology (e.g., type of study performed, aim of study, duration of study), participant information (e.g., population description, recruitment method, age, sex, comorbidities, subgroups), outcomes (e.g., outcome description, diagnostic criteria of outcome, risk estimates, power, sample size), data analysis, and key conclusions.

The searches from MEDLINE, EMBASE, Agricultural & Environmental Science Collection, CAB Abstracts, and Scopus databases, yielded 6,154 articles that led to screening of 6,125 titles and abstracts after the removal of duplicates (Fig. 1). A total of 44 articles were identified for full-text screening. Twenty articles were identified as relevant. Of these, 11 articles were excluded because they lacked mention of organophosphates. An additional reference was identified from the reference list of an article during full-text review. Ultimately, a total of 10 articles were included in this scoping review (Centers for Disease Control and Prevention 2003, 2005, O’Sullivan et al. 2005, Peterson et al. 2006, Duprey et al. 2008, Scharf et al. 2008, Valcke et al. 2008, Schleier et al. 2009, 2015, Likos et al. 2016).

Of the 10 included studies, only two articles were published within the last 11 years (1 January 2013–22 May 2024) (Schleier et al. 2015, Likos et al. 2016), and eight were published from 2000 to 2012 (Centers for Disease Control and Prevention 2003, 2005, O’Sullivan et al. 2005, Peterson et al. 2006, Duprey et al. 2008, Scharf et al. 2008, Valcke et al. 2008, Schleier et al. 2009).

The majority of analyzed studies were conducted within the United States (n = 9) (Centers for Disease Control and Prevention 2003, 2005, O’Sullivan et al. 2005, Peterson et al. 2006, Duprey et al. 2008, Scharf et al. 2008, Schleier et al. 2009, 2015, Likos et al. 2016); one study was conducted in Canada (Valcke et al. 2008). Of the studies conducted in the United States, data were included from Arizona, California, Florida, Louisiana, Michigan, New York, North Carolina, Oregon, Texas, Virginia, and Washington. Two separate studies reported data collected from multiple states (Centers for Disease Control and Prevention 2003, Likos et al. 2016). All other included studies only reported data from a single geographical location. Weather data from New York and Massachusetts were used for two risk assessment studies.

All included studies had defined methods of insecticide application: ground application, or aerial spraying by helicopter or airplane. The ULV spraying methods were specified in eight studies (O’Sullivan et al. 2005, Peterson et al. 2006, Valcke et al. 2008, Schleier et al. 2009, 2015); two studies did not specify ULV methods but conducted spraying events in accordance with insecticide labeling and regulations from the U.S. EPA and Health Canada Pest Management Regulatory Agency (PMRA) (Centers for Disease Control and Prevention 2003, Scharf et al. 2008, Environmental Protection Agency 2023a). One study used absorption of aqueous malathion as a proxy to measure potential insecticide exposure from recreational swimming during aerial spraying (Scharf et al. 2008). Malathion and naled were used commonly for adult mosquito control and were mentioned in all included articles. Only two studies examined health effects associated with a single insecticide (e.g., either only malathion or only naled) (27, 31). The other eight studies examined insecticide exposures to a combination of malathion, naled, and other insecticides that were not organophosphates (Centers for Disease Control and Prevention 2003, 2005, O’Sullivan et al. 2005, Peterson et al. 2006, Duprey et al. 2008, Scharf et al. 2008, Valcke et al. 2008, Schleier et al. 2009, 2015, Likos et al. 2016).

Three types of study design were represented in the included studies: cohort (n = 5), case study (n = 1), and risk assessment (n = 4). The five cohort studies included both retrospective and prospective cohort studies.

In Table 2, we summarize five articles describing the cohort studies evaluating human health effects as an outcome of organophosphate insecticide exposure (Centers for Disease Control and Prevention 2003, 2005, O’Sullivan et al. 2005, Duprey et al. 2008, Likos et al. 2016). Two studies were retrospective investigations that examined the severity of insecticide-related illness and injury and hospitalizations associated with asthma exacerbation following insecticide exposure (Centers for Disease Control and Prevention 2003, O’Sullivan et al. 2005). One cohort monitored exposure calls and emergency department (ED) visits associated with insecticide exposure before and after a spray event in 2016 (Likos et al. 2016). This study also compared historical data with data collected during the 2016 spray event for ED visits for asthma, reactive airway disease, wheezing, or shortness of breath. Two prospective cohort studies assessed urine levels of a nonspecific organophosphate metabolite, dimethyl phosphate (DMP), before and after a spray event (31, 32). Participants of each of the cohort studies, except for Duprey et al. and Centers for Disease Control and Prevention (CDC), were recruited through hospitalization records, poison information call center records, or state-based surveillance of acute insecticide-related illness and injury cases that matched residential zip codes with spray events (Centers for Disease Control and Prevention 2005, Duprey et al. 2008). Duprey et al. recruited participants through random households chosen within each proposed spray zone based on sample clustering (Duprey et al. 2008). The CDC study published in 2005 recruited individuals randomly chosen within census blocks marked for spraying (Centers for Disease Control and Prevention 2005). All participants in each of the cohort studies were living in residential areas exposed to spray events and ranged in age and sex from adult females (aged 18–76 years), adult males (18–76 years), youth (10–12 years), children (5–6 years), toddlers (2–3 years), and infants (0.5–1.5 years). No data were reported for children aged 1.5–2 years.

In Table 3 we summarize the single case study. Scharf et al. examined the dermal absorption rate of malathion and conducted a simulation experiment to determine the potential health impact on recreational swimmers exposed to waters following a malathion application (Scharf et al. 2008). In this study, 20 volunteers (n = 2 female, n = 18 male) aged 25–50 years had dermal exposure to aqueous malathion and absorption rates were compared to a hazard index (HI). This study was used to extrapolate information to determine theoretical insecticide values that recreational swimmers might be exposed to after an aerial spraying event. In this single study, exposure to aqueous malathion in concentrations that could be expected during recreational swimming following an aerial spray event did not result in substantial dermal absorption in simulated test subjects.

In Table 4 we summarize four risk assessment studies (Peterson et al. 2006, Valcke et al. 2008, Schleier et al. 2009, 2015). Toxicological risk is often expressed as risk quotients (RQ), which are determined through simulated environmental concentrations, exposure routes, and known toxicity endpoints; RQ values below 1.0 signify levels not warranting regulatory concern (Environmental Protection Agency 2023b). The exposure pathways range from acute to subchronic exposures, encompassing various scenarios such as inhalation, dermal exposure from spray deposition, hand-to-mouth exposure, and ingestion of garden produce, among others.

The four referenced studies assess risk to human health linked to insecticide use in mosquito control, focusing on different exposure routes and demographics (Peterson et al. 2006, Valcke et al. 2008, Schleier et al. 2009, 2015). They employed Bayesian inference for RQ determination relating to health effects from organophosphate exposure. Setting it apart, the study presented in Schleier et al. 2015 is a meta-analysis that includes data from the other three studies (Schleier et al. 2015). Schleier et al. 2015 revised an earlier RQ estimation for adult males exposed to malathion from 0.02 to 0.41, integrating new insights from additional studies (Peterson et al. 2006, Valcke et al. 2008, Schleier et al. 2009, 2015).

The remaining three studies covered more than one insecticide but included malathion and naled. The studies analyzed exposure from both ground or aerial in silico ULV spraying, using the AERMOD/ISCST3 or AGDISP models that assessed for the potential of drift under specified atmospheric and operating conditions (Bilanin AJ et al. 1989, Teske ME et al. 2019), and assessed the risk across a range of ages (Table 4) (Peterson et al. 2006, Valcke et al. 2008, Schleier et al. 2009). Assumptions made for the ULV simulations and risk estimates were broadly comparable and included: (i) all chemicals had a 24-h half-life in air; (ii) naled had an 18-h half-life; (iii) insecticides were applied at maximum application rates; and (iv) all insecticides were subject to the same weather conditions with no weather stress during the spray event.

Peterson et al. (19) indicated that aerial applications resulted in significantly lower deposition levels, denoting reduced human exposure and risk (Peterson et al. 2006). This contrasts with Valcke et al. (2008), who found increased risks with aerial spraying, likely due to increased application rates (Valcke et al. 2008). Notably, Peterson et al. did not publish their findings on aerial ULV (Peterson et al. 2006). The low risk assessed from aerial spraying found by Peterson is likely the reason why Schleier et al. (2009) only evaluated ground ULV risk.

Contrasting ground applications, Peterson and Valcke reported significantly different acute air concentrations at similar application rates (9.76 μg/m³ vs. 159.6 μg/m³ for 61 g/ha and 63 g/ha, respectively) (Peterson et al. 2006, Valcke et al. 2008). This difference in air concentrations likely influenced RQs values, although final risk assessment requires consideration of more factors than application rates.

In Fig. 2 some reported RQ values for each population subgroup and exposure routes are outlined for the included studies. With the exception of acute and subchronic aerial exposure for infants and toddlers reported in Valcke et al. (2008), all other reported RQ’s were less than 1.0 (Peterson et al. 2006, Schleier et al. 2009, 2015). The RQ values >1.0 may indicate possible adverse health effects. Unfortunately, methodological differences (19, 24, 25, 26) make direct comparison among the studies challenging. Furthermore, age is used as a proxy for body weight and exposure behaviors, and the age categories (infant, toddler, etc.) are not the same.

This analysis is subject to several limitations. First, considering that organophosphates are commonly used for vector control, we only identified 10 studies relating to exposure to organophosphates with reported human health effects. Within those 10 studies, only two organophosphates were represented, malathion and naled. Other organophosphates (i.e., chlorpyrifos) have historically been used; however, given changes in organophosphate manufacturing and subsequent use, malathion and naled are the organophosphates used predominately for adult mosquito control. Secondly, the studies included in this review could not determine causality between exposure to adulticides and development of illness. Insecticide exposure often occurs alongside exposure to other chemicals or environmental factors that can influence health outcomes. Failing to account for these potential confounding factors can obscure the true relationship between insecticides and health effects. Finally, we only considered published articles without assessing the existing evidence in grey literature and other sources. The nonconsideration of grey literature and other sources restricts our findings to what was reported by scientific journals and identified in this literature search and could possibly prevent analysis of relevant cases that were rejected for publication or reported in other forums.