Using systematic review methods, this scoping review of publicly available literature indexed in PubMed identified a heterogenous body of evidence on potential health effects data associated with neonicotinoid exposures. The interactive evidence map allows researchers to sort and explore the literature by pesticide, broad health effect categories, and evidence stream (Figure 2). Although 191 studies were identified as potentially relevant to the human health effects of neonicotinoids, the studies varied considerably by endpoints examined and evidence stream as well as by study design, including human observational studies, case series and case reports in humans and animals, and dozens of experimental studies in animals, including rodents, primates, fish, C. elegans, and Drosophila. The clear majority of the research has been performed on the most widely used neonicotinoid—imidacloprid (127 publications)—with as few as four publications identified for each of the other six chemicals (n = 4–34 publications). Neonicotinoids are insecticides that are neurotoxic to insects through insect nicotinic acetylcholine receptors (nAChRs) and, as might be expected from this mechanism, neurological effects were the most-studied health effects across all evidence streams including an array of endpoints measured in experimental animal studies (Figure 3).

Twenty-five studies of humans mentioned neonicotinoid exposure in relation to specific health effects. The majority of the human evidence is limited to case reports (n = 19), which covered a range of observed human health effects from neurological effects to hematological and cardiovascular effects after acute, high-exposure scenarios. In addition to the case reports, there were five case-control studies and one cross-sectional study (Table 2). While four of the case-control studies evaluated congenital/developmental effects, they all evaluated different developmental effects, including various birth defects, such as congenital heart defects (Carmichael et al. 2014; Keil et al. 2014; Shaw et al. 2014; Yang et al. 2014), gastroschisis (Shaw et al. 2014), and anencephaly, cleft lip, and spina bifida (Yang et al. 2014). One measured autism spectrum disorder, which was also considered a neurological effect (Keil et al. 2014). The remaining case-control and cross-sectional studies reported neurological symptoms and biochemical measures (Khan et al. 2010; Marfo et al. 2015). The limited number of epidemiological studies is a critical data gap for assessing the potential health effects of neonicotinoids, as described in more detail in a previously published systematic review (Cimino et al. 2017). Additional high-quality studies on the same or related endpoints would help develop bodies of evidence for reaching conclusions on the potential association between exposure to neonicotinoids and any specific health effect.

All animal models were considered relevant to this review if used to investigate mechanisms potentially relevant to human health (e.g., Drosophila studies). However, studies focused solely on the efficacy of pesticides to insect pests or those that described effects on nontarget insects were not included in this review. Using these criteria, a relatively large body of animal evidence was discovered with 86 studies identified that used nonhuman mammalian models, 25 that used fish, and five using C. elegans or Drosophila (Figure 2). Despite these relatively high numbers, considerable heterogeneity was observed in study designs and investigated health effects. The largest pocket of studies within a related health category was for neurological outcomes (Figure 3). Although 42 of the studies evaluated neurological effects (one of which was a case series in three dogs), the endpoints varied with few covering the same category of neurological effect (e.g., anxiety, motor function, or learning and memory) or specific endpoint measured (e.g., acetylcholinesterase, escape time in the Morris maze, flight path in bats).

Although 29 in vitro studies were also relevant to neurological or developmental effects, 11 of these studied the effects on mammalian nAChR only (Table 3), and too few studies investigating other similar endpoints were available to evaluate any additional potential specific mechanism or mode of action for biological plausibility using in vitro data (Table 4). The nAChR-dependent mechanism by which neonicotinoids exert effects on target organisms (insects) is well established; however, interactions of neonicotinoids with human receptors are less clear. Case reports of acute poisonings have described nicotinic-like symptoms (see human case reports in Figure 2), and decreased plasma cholinesterase levels and nicotinic symptoms were reported in subjects with higher measured neonicotinoid levels in plasma or urine, respectively (Table 2) (Khan et al. 2010; Marfo et al. 2015). However, these data are limited to a small number of exposed cases outside of the United States and included acute exposures and co-exposures to multiple pesticides not limited to neonicotinoids.

In addition to the studies considered relevant for human health effects, a large body of evidence was also available from studies on neonicotinoids that did not meet the study eligibility criteria due to the lack of endpoints directly relevant to human health. The majority of excluded neonicotinoid studies that did not report human health effects-related endpoints reported target insect responses and included efficacy studies in pets (e.g., killing fleas on pets). A large proportion of studies evaluated ecological effects including effects on off-target insects or other invertebrates, such as honeybees, which are beyond the scope of this review of human health effects as their nAChRs are distinct from mammals. However, a few studies focused on exposure and/or pharmacokinetics (i.e., absorption, distribution, metabolism, or excretion in human, animal, or in vitro studies) for neonicotinoids that could be useful when evaluating human health effects.