Widespread occurrence of pesticides in low-income housing

The study addresses gaps in knowledge about indoor pesticide exposures in non-agricultural settings, particularly among low-SES residents in social housing who face elevated risk of pest infestations and frequent pesticide applications. Children and pregnant women are especially vulnerable. While most research has focused on agricultural exposures, indoor residential pesticide use (for vermin control, treated consumer products and building materials, and tobacco contamination) can lead to persistent indoor reservoirs and chronic exposure. Regulatory actions have phased out many organochlorine (OCP), organophosphate (OPP), and carbamate pesticides, shifting residential use toward pyrethroids/pyrethrins (PYRs). The research question is to quantify indoor air concentrations (and inferred exposures) to legacy and current-use pesticides in Toronto social housing and explore determinants such as smoking and building-level patterns.

Prior literature shows higher pesticide exposures in low-income or poor-quality housing, with persistence of legacy OCPs linked to building age and historical use. Children exhibit higher biomarker levels for PYR metabolites; maternal-fetal transfer of contemporary-use pesticides has been reported. Restrictions under the Stockholm Convention and national regulations have reduced or banned many OCPs and limited indoor use of OPPs/carbamates; PYR use has increased, though emerging evidence links PYRs to neurodevelopmental and reproductive effects. Indoor persistence of semi-volatile pesticides is well documented, with associations to adverse health outcomes, and few Canadian data exist for indoor pesticide air levels outside legacy OCPs.

Study setting: Seven social housing multi-unit residential buildings (MURBs) from the 1970s at three Toronto sites, each ≥65 km from agricultural areas. Units categorized as seniors, families, or bachelors. Sampling: Forty-six units sampled in winter 2017. Each unit deployed an Amaircare XR-100 portable air cleaner with 127 mm high-efficiency media for one week (median flow 39.2 m³/day). Devices were placed ~30 cm from the ceiling to avoid tampering. Surveys on resident behaviors (e.g., smoking) and housing characteristics were conducted during 2015–2017 visits. Target analytes: 28 pesticides (legacy and current-use) including OCPs (e.g., DDT, heptachlor, lindane, endosulfans), OPPs (malathion, trichlorfon, diazinon), PYRs/pyrethrin (e.g., pyrethrin I, permethrin, allethrin, tetramethrin, L-cyhalothrin, prallethrin, cyfluthrin), strobilurins (azoxystrobin, fluoxastrobin, trifloxystrobin), and others (imidacloprid, propiconazole, pendimethalin). Analytical methods: Half of each filter was analyzed. Samples and blanks were spiked with six labeled surrogates; five field blanks and nine lab blanks were included. Extraction via sonication three times with hexane:acetone:dichloromethane (2:1:1, 20 min each), combining supernatants, reducing under N2, Florisil SPE cleanup, final concentration to 0.5 mL, mirex internal standard addition. Analysis by GC-MSD (Agilent 7890B/5977A) in EI and CI modes. QA/QC: Surrogate recoveries 62–83% (results recovery-corrected). Blank correction criteria applied (no correction <5% of sample; correct at 5–35%; discard >35%). MDLs from mean lab blank + 3 SD or instrument LOD by S/N ≈10:1 if absent in blanks. Blank levels were <MDL or very low; values <MDL not substituted due to DF ≤50%. Concentration calculation: Particle-phase air concentration derived from filter PM mass, analyte concentration in PM, filtration efficiency (assumed 100%), device flow rate, and sampling time. Gas-phase estimation: Using Harner–Bidleman partitioning with Koa (OPERA 2.6) and fom=0.4 to compute Kp, then Cg = Cp/Kp; total air = particle + estimated gas-phase. Data analysis: Normality assessed (Chi-square) indicated non-normal distributions; nonparametric tests used. Spearman rank correlations assessed co-occurrence; Mann–Whitney Wilcoxon compared smoking vs non-smoking units (p<0.05). Analyses done in Excel 365 and R 4.1.2.

• Scope and detection: 24 of 28 target pesticides detected; at least one pesticide detected in 89% of units. Detection frequencies (DF): OCPs 0–50%; OPPs 11–24%; PYRs 7–48%; strobilurins 7–22%; imidacloprid 22%; propiconazole 15%; pendimethalin 41%. - Pyrethroids/pyrethrin: ΣPYRs particle-phase maximum 36,000 pg/m³. Pyrethrin I had DF 48% and the highest particle-phase concentration of all compounds (32,000 pg/m³). Allethrin (max 16,000 pg/m³), permethrin (max 14,000 pg/m³), L-cyhalothrin (max 6,000 pg/m³). Estimated total air ΣPYRs maximum 740,000 pg/m³ (range using EPISuite Koa 110,000–270,000 pg/m³). Allethrin and permethrin concentrations exceeded most prior low-SES indoor air studies, though some New York data for permethrin were higher. - Organochlorines: ΣOCPs particle-phase maximum 4,400 pg/m³. Heptachlor (restricted 1985) particle-phase max 2,600 pg/m³ and p,p'-DDT (restricted 1985) max 1,400 pg/m³; chlorothalonil max 1,200 pg/m³ (DF 50% despite 2011 domestic-use suspension). Lindane max 990 pg/m³. Estimated total air maximum for heptachlor was 443,000 pg/m³; several OCPs (heptachlor, lindane, chlorothalonil, endosulfan I) were 2–11× higher than reported in earlier residential studies. - Organophosphates: Σ3OPPs (malathion, trichlorfon, diazinon) particle-phase max 3,600 pg/m³; trichlorfon max 3,600 pg/m³; malathion max 2,800 pg/m³. Estimated total air Σ3OPPs max 77,000 pg/m³ (EPISuite Koa range 60,000–200,000 pg/m³). Diazinon and malathion often higher than some earlier low-SES studies; lower than some New York/Northern California results. - Strobilurins: ΣSTRs particle-phase max 1,200 pg/m³; estimated total air max 1,300 pg/m³. First reported indoor air concentrations for STRs; likely from treated building materials. - Other pesticides: Imidacloprid particle-phase max 930 pg/m³; estimated total air max 34,000 pg/m³. Propiconazole particle-phase max 1,100 pg/m³; total air max 2,200 pg/m³. Pendimethalin particle-phase max 4,400 pg/m³; total air max 9,100 pg/m³. - Correlations/co-occurrence: Significant correlations between p,p'-DDT and p,p'-DDE; endosulfan I and II. Co-occurrences consistent with co-formulations (e.g., pyriproxyfen with tetramethrin), and frequent co-use signals among PYRs (permethrin with pyrethrin I: co-occurred in 16 units; pyrethrin I with allethrin: 7 units). Correlations among pendimethalin, permethrin, and chlorothalonil (tobacco-related). Additional associations observed between permethrin and strobilurins. - Smoking association: 30% of units reported smoking. Five pesticides—chlorothalonil, permethrin, pyriproxyfen, pyrethrin I, pendimethalin—had DF >60% in smoking units and significantly higher concentrations than in non-smoking units (Mann–Whitney, p<0.05), aligning with pesticides used on tobacco crops. - Building-level patterns: Pesticide profiles and concentrations clustered by building, independent of unit type, suggesting building-wide eradication activities and/or inter-unit transfer for more volatile compounds.

The findings demonstrate widespread indoor exposure to both legacy and current-use pesticides among residents of low-income social housing, addressing a critical data gap for Canada. Despite historical bans and restrictions, legacy OCPs (e.g., heptachlor, DDT, lindane) persist indoors at detectable levels with high estimated total air concentrations, highlighting prolonged indoor reservoirs and limited degradation. Current-use pyrethroids/pyrethrins exhibited the highest detection frequencies and particle-phase concentrations, consistent with modern residential pest-control practices and treatment of head lice/pets, particularly relevant in settings with higher pest burdens. Significant correlations and co-occurrences indicate co-formulation/co-use patterns and multiple pesticide applications in units. The association between smoking and elevated levels of five pesticides linked to tobacco production suggests tobacco smoke as an additional indoor source, with potential migration into non-smoking units. Building-specific clustering further implies centralized eradication activities or shared sources within MURBs. Collectively, the results underscore the importance of considering multiple sources (intentional pest control, treated materials, tobacco), building-level practices, and the persistence of banned pesticides when assessing exposure risks in vulnerable populations.

In 46 social housing units in Toronto, at least one of 28 target pesticides was detected in 89% of homes. Current-use pyrethroids/pyrethrins had the highest detection frequencies and particle-phase concentrations, while legacy OCPs—despite long-standing restrictions—had very high estimated total air concentrations, indicating strong indoor persistence. Several pesticides not registered for indoor domestic use (e.g., strobilurins) were detected, likely from treated building materials. Five pesticides associated with tobacco crops were significantly elevated in smoking units. These results provide the first comprehensive Canadian indoor air dataset for many pesticides and indicate widespread, multi-pesticide exposure among low-SES social housing residents. Potential future research should include collecting records of pesticide applications by building management and residents, direct gas-phase measurements and personal exposure monitoring, and comparative studies across SES groups and building types to contextualize exposures and inform interventions.

• No records of pesticide application by building management or residents, limiting source attribution. - Lack of a comparison group from higher-SES households due to limited available data and resources. - Air sampling was conducted near the ceiling, away from the breathing zone, which may affect representativeness. - Sampling captured particle-phase directly; gas-phase concentrations were estimated via partitioning assumptions rather than measured directly.