Changes in Indoor Insecticide Residue Levels after Adopting an Integrated Pest Management Program to Control German Cockroach Infestations in an Apartment Building

Insecticide use in homes leads to human exposure to insecticide residues that persist in the environment. Integrated pest management (IPM) programs have been known to be more environmentally friendly for managing German cockroach (Blattella germanica L.) infestations, but their effect on indoor insecticide residue levels are not well understood. An IPM program consisting of applying cockroach gel baits and placing insect sticky traps as the primary treatment methods were implemented. Floor wipe samples were collected from the bedroom and kitchen floors of 69 apartments with German cockroach infestations at 0 months and again at 12 months from 49 of the 69 apartments sampled at 0 months. Levels of 18 insecticide residues were measured. The mean insecticide residue concentration per apartment decreased by 74% after 12 months. The number of insecticides detected per apartment decreased from 2.5 ± 0.2 to 1.5 ± 0.2 (mean ± standard error). Indoxacarb residue was only detected in two apartments at 12 months despite the fact that an average of 32 ± 4 g 0.6% indoxacarb gel bait was applied per apartment. IPM implementation can result in significant reduction in the insecticide residue concentrations and number of detected insecticides in floor dust samples.


Quality Assurance/ Quality Control
All samples in this study were spiked with two surrogates to monitor both the extraction and analysis efficiency. Dibutyl chlorendate and triphenyl phosphate were the two surrogates added to each sample at approximately 12 µg each. Laboratory blanks and matrix blanks were analyzed along with the surface wipe samples as controls. Sample spikes consisting of all the insecticide analytes at eight different concentration levels throughout the working range were used to determine the recovery of each insecticide. The insecticide neat standards were obtained from the EPA pesticide repository located in Fort Meade Maryland. Purity of the analytical standards are shown in Table S1. Stock standards were prepared from the neat standards at a concentration of approximately 400 ug/ml in ethyl acetate. Intermediate insecticide standard mixtures were prepared from the individual stock standards at a concentration of approximately 40 ug/ml for each insecticide analyte. The amounts added for each insecticide in the eight concentration levels were as follows: 0.2 µg, 0.4 µg, 0.6 µg, 0.8 µg, 1.0 µg, 2.0 µg, 4.0 µg, 5.0 µg. Since pyrethrin is a multi-component insecticide analyte, higher spike concentrations were used. The eight concentration levels for pyrethrin were as follows: 1.0 µg, 2.0 µg, 3.0 µg, 4.0 µg, 5.0 µg, 10 µg, 20 µg, 25 µg. The average recovery and relative standard deviation for each insecticide analyte based on thirty-two spiked samples can be found in Table S2. The average recovery and relative standard deviation for the two surrogates which were based on 309 samples can also be found in Table S2.
The sample analysis utilized an external standard technique with calibration curves consisting of six concentration points for insecticide analyte and a linear regression algorithm was used for quantification. The method and instrument detection limits are also listed for each insecticide analyte in the table above. EPA pesticide external standard methods for water samples such as method 608 was used as a model for the analysis of the swab samples.
The extraction procedure is like surface wipe extractions performed by Federal Insecticide Fungicide Rodenticide Act (FIFRA) state laboratories for pesticide enforcement cases. The United States Environmental Protection Agency Office of Enforcement and Compliance Monitoring generated the "NEIC Pesticide Sampling Guide" which was used as a model for this study. This guide was generated in August 1985 by Robert F. Schneider for the National Enforcement Investigations Center which was in Denver, Colorado.
Qualitative identification of insecticide analytes in samples were confirmed when the relative abundance of characteristic ions in the mass spectrum of both the standard and the sample agreed within 20% absolute abundance. For example, if a characteristic ion has a relative abundance of 30% in the standard spectrum, its abundance in the sample spectrum must be in the range of 10 -50%. In this study there were no qualitative identification of insecticide analytes below the method detection limits (MDL) and if insecticide analytes were qualitatively confirmed below the MDL the quantified value would be reported as the MDL since quantifying in this region would be beyond the linear range.
The MDL was determined by spiking eight blank swab samples with approximately 40 ng of each insecticide analyte for a final concentration of approximately 0.22 ng/cm 2 . The eight spiked samples were analyzed and the standard deviation was determined for each insecticide analyte. Since there were 7 degrees of freedom from the student t test the standard deviation for each insecticide analyte were multiplied by 2.998 to generate the MDL values. The instrument detection limits were determined for each insecticide analyte by obtaining the peak height of three times the signal to noise level and comparing it to the peak height of the lowest concentration data point in the calibration curve to generate the value.
The column pressure setpoint was adjusted to 6894.7 pascals at a constant flow rate setting to produce a linear velocity of 100 cm/sec and a flow rate of 13.2 ml/min. The injector temperature was set to 523ºK and a splitless injection was used with a split vent of 50ml/min at 0.7min. The mass spectrometer ion source was set at 503ºK and the temperature of the quadrupole was 423ºK. The transfer line leading to the mass spectrometer was set at 883ºK. Elution of the insecticide analytes from the GC column occurred through the temperature program that ranged from 323ºK to 573ºK over a 7-minute chromatographic run. * Paired student t test was used for data that are normally distributed. Wilcoxon signed rank test was used for data that are not normally distributed. Non-normally distributed data were not transformed before analysis.