Pesticides were detected in all studied fruits and vegetables, with 39 active ingredients (AIs) detected in all samples and 18 AIs in at least some of the food samples. Fonofos, fenitrothion and fenhexamid concentrations were above the MRLs in watermelon, passion fruit, tomato, cabbage and eggplant. Risk assessment calculations show that EDIs for 18 pesticides were above the ADI in some cases, with HQs that ranged from 1 up to 443 and thus may pose chronic health risks. Children experienced the highest HQs and therefore potentially higher chronic health risks from pesticide residues in fruits and vegetables.
Overall, 29% of the pesticides we tested for had EDIs over an ADI. This is a high proportion of exceedances compared to other risk assessment studies [46–48]. When calculated by stage along the supply chain and age group, 16 and 18 pesticides respectively had high EDIs are above their ADI. As discussed by JA Vaccaro and FG Huffman [49], age is a key dietary risk factor that should be considered while performing health risk assessment Several fruit and vegetable surveillance studies have estimated EDI and similar EDIs. Studies in Chile, Poland and Kazakhstan had EDIs ranging from < 0.001 to 5.2 [47, 50–54], which is within the range of our findings.
Many pesticides were detected in all studied fruits and vegetables with levels below the EU MRLs except for Fonofos, fenitrothion and fenhexamid. Our findings are consistent with existing literature showing detection of many pesticides in fruits and vegetables [50, 51, 55, 56]. Like our findings, many past studies have pesticide residue levels that are above MRL values, especially organophosphates like fenitrothion [46, 48, 53, 54, 57–60]. For example, recent studies in Ghana and Nigeria also found that many pesticides residue levels in produce were above the respective MRLs [42, 61]. The most frequently detected pesticides that have exceeded MRLs have been organophosphates, carbamates, pyrethroids and neonicotinoids based on studies in Uganda, Ghana, Egypt, Poland and Chile [4, 5, 28, 34, 42, 43, 46, 62], especially in leafy vegetables [50, 51]. Given that MRLs are determined based on good agricultural practices (GAPs) in field experiments and not necessarily health risks [63, 64], consumption of pesticides below the MRLs might exceed health-based exposure benchmarks depending on individual consumption patterns.
Our findings confirm similar findings to other studies carried out in Poland, Nigeria and Saudi Arabia which found that many pesticides had a HQ > 1 [46, 65, 66]. On the other hand, literature from Turkey, Poland, Ghana, China and South Korea showed no chronic health risk associated with pesticide residues in fruits and vegetables [53, 54, 58, 60, 67–70]. using probabilistic modelling, Z Eslami, V Mahdavi and B Tajdar-Oranj [57] in Iran found that pesticide residues did not pose health risks to adults and children. When assessed by stage along the supply chain, some pesticide showed a low HQ and consequently lower risk when consumed at farm than at other stages further along the supply chain, such as restaurants and homes. Our findings are similar to those from previous studies which have shown a higher chronic health risk for stages upstream along the chain [62, 71]. When HQ was assessed by age, children more frequently experienced higher hazard quotients (18 − 13) compared with adults (11 − 9) with HQS up to 443, compared with a maximum HQ for adults at XX. Our findings are similar to findings from studies from Chile, Nigeria and China that assessed risk by age which found that chronic health risks were higher in children compared to adults [50, 53, 69, 72].
Our findings have implications on policy and future research. We used the EU MRLs and ADIs to evaluate exposures and risks, these benchmarks are lower and hence more sensitive than other guidelines. For example, Codex Alimentarius guidelines are higher, which would suggest lower health risks based on the exposure we evaluated. There is a need to develop Ugandan standards for MRLs and ADI based on local studies and context. The high HQs demonstrate in our study also demonstrate the need for routine monitoring and surveillance of pesticide residues in foods, especially in fruits and vegetables.
This study has several strengths and limitations. This study is the largest in Uganda to examine pesticide residues in fruits and vegetables; and we interviewed over 2000 residents to obtain information on dietary intake patterns. Dietary consumption data for fruit and vegetable was measured using a contextualised food album and thus presents a true reflection of the study community. We used mean residue concentrations to assess likely average exposures to consumers, but individual variability in eating patterns may result in higher or lower chronic exposures [73]. Additionally, we computed hazard quotients for consumption of individual foods. It is likely that consumers ate several different fruits or vegetables on any given day. In future analyses, we will use probabilistic methods to assess the range of potential exposures and health risks from more realistic diet patterns. We will also apply relative potency factors (RPFs) to assess cumulative health risks for pesticide classes with established RPFS [74]. Fruits and vegetables were not tracked from farm to fork during sampling due cost and time challenges. Future studies examining pesticide residues along the farm to fork chain should track and sample individual produce lots from harvest to the consumer. Additionally, this study was carried out in a primarily urban community and may not represent a typical Ugandan rural setting. Finally, dietary consumption measurement did not cover the broad spectrum of fruits and vegetables but rather focussed on commonly consumed items within the study area (watermelon, passion fruit, tomato, cabbage and eggplant). However, the study area represents a large proportion of the Ugandan population and several commonly eaten foods.