To capture opossums, we deployed Havahart 1-door cage traps (81 cm × 31 cm × 25 cm) (Woodstream Corporation, https://www.woodstream.com) in Galveston, Texas, USA, during evening hours. We conducted 2 arms, comprising an experimental group and a control group. In the experimental group, we baited traps with canned cat food mixed with a crushed 270-mg spinosad tablet. In the control group, we trapped opossums using cat food without spinosad in the bait.

Figure

Figure. Trapping and flea collection in a study of use of oral flea preventive to control Rickettsia typhi–infected fleas on reservoir opossums, Galveston, Texas, USA, 2023–2024. A) Trapped opossum;...

Video

Video. Video of opossum trapping in a study of use of oral flea preventive to control Rickettsia typhi–infected fleas on reservoir opossums, Galveston, Texas, USA, 2023–2024. A cage trap...

Spinosad starts to kill fleas within 30 minutes, and studies show it can kill 100% of fleas within 4 hours of ingestion in dogs and within 24 hours in cats (4). To collect dead fleas fallen from opossums, we placed a white terrycloth towel (160 cm × 84 cm) under the trap (Figure; Video). To ensure trapped opossums ingested bait at least 2 hours before flea collection, we checked traps between 4 am and 6 am the morning after setting.

We used intramuscular ketamine (30 mg/kg) to anesthetize trapped opossums. After adequate sedation, we collected fleas by combing through the fur with a flea comb (Hartz, https://www.hartz.com) for 15 minutes, then placed fleas on masking tape and secured in sealable plastic storage bags (Figure, panels B–D). After we collected fleas from the anesthetized opossum, we collected fleas from the towel. During collections, we observed flea activity and considered moving fleas to be alive and fleas without movement to be dead. All opossums recovered uneventfully from anesthesia and were released at a distant site to avoid subsequent trapping of the same animal. We caught 3 off-target species (1 raccoon and 2 cats) and released those animals on-site without intervening procedures.

We captured 9 opossums over 19 trap nights: 5 in the experimental group and 4 in control group (Table 1). Six (66.7%) were female and 3 (33.3%) were male. The mean weight was 2.1 kg. All opossums completely ingested food within the trap. We observed fleas on all 9 opossums. Among the opossums that ingested spinosad, fleas were largely immobile and easily combed from fur; fleas were also numerous on the towel and were easily collected and bagged (Figure, panels C, D). In contrast, fleas in the control group moved rapidly within the opossums’ fur; were difficult to comb off, often jumping off the comb; and would frequently escape from the tape after being bagged.

We examined collected fleas with a dissecting microscope and identified flea species by taxonomic key (5). We pooled fleas (dead and alive) in groups of 5–20 fleas based on the animal from which they were collected. We homogenized a subset of pools from each opossum and extracted DNA, as previously described (6). We used a multiplex real-time PCR to amplify and identify R. typhi and R. felis from flea homogenates (7).

We compared the proportion of dead fleas from the experimental and control groups. Because those proportions were bounded by 0 and 1, we transformed proportions by taking the arcsine of their square root (8). We needed sample sizes of at least 3 opossums in the experimental and 2 in the control groups to achieve 82% power to reject the null hypothesis with a significance level (α) of 0.05 using a 2-sided 2-sample equal-variance t test. We determined rates of R. typhi and R. felis flea infection by using previously described methods (9) (Appendix).

We collected a total of 2,047 fleas and identified all as Ct. felis. The proportion of dead fleas from opossums in the experimental group was 0.98 (range 0.92–1) compared with 0 (p<0.001) in the control group (Table 1).

We tested 71 pools (comprising 744 fleas) by PCR; 11 pools demonstrated R. typhi DNA, and 23 pools demonstrated R. felis DNA (Table 2). Those findings correspond to an infection rate of 1.3% (95% CI 0.7–2.2) for R. typhi and 3.2% (95% CI 2.2–7.9) for R. felis. Sequenced PCR products of portions of rickettsial htrA (GenBank accession no. PQ625781) and sca5 (GenBank accession no. PQ625780) of an R. typhi–infected flea pool detected by real-time PCR confirmed 100% homology with R. typhi Wilmington type strain.

In this study, we showed that commercially available spinosad (approximately $20 per tablet) effectively kills fleas infesting opossums. The observed animals readily fed on bait laced with the medication, and almost 100% of collected fleas were found dead at least 2 hours after opossums ingested the preventive. Although spinosad caused mild side effects (e.g., vomiting) in dogs and cats in controlled studies (4), little data on adverse effects in other species are available. Although we noticed no ill effects in opossums this study, the sample size was small.

Consistent with other areas of the United States where murine typhus is endemic, we detected R. typhi in cat fleas collected from opossums (6,10). In contrast, other reports show few fleas collected from domestic cats are infected with R. typhi (11,12).

Targeted efforts to control rat fleas via application of DDT on rat runs and harborages during the 1940s and 1950s exemplify how vector control can break the cycle of vectorborne disease transmission. In 1944, murine typhus peaked at 5,401 reported cases, but that number fell to <100 reported cases annually by the mid-1950s after introduction of DDT (2,3). However, the effects of DDT seemed to have little spillover into opossums (13), and the lingering low-level endemicity of murine typhus in the United States was eventually attributed to an alternate R. typhi transmission cycle involving opossums and cat fleas (1,3). Those findings support the need to focus flea control efforts on opossums in murine typhus endemic areas of the United States. More research is required to investigate the proposed methods’ feasibility, effectiveness in nonexperimental conditions, logistics and toxicity among nontarget species, and effects on disease transmission in murine typhus endemic areas. For instance, in high-burden disease locations, field-deployable methods could include manual dispersion of spinosad-impregnated food pellets or timed-release dispenser stations placed at strategic distances.

In conclusion, the growing distribution and incidence of murine typhus in the United States are a call to develop and implement integrated pest control strategies to aid public health prevention (14,15). We found that oral flea preventive effectively controlled R. typhi flea vectors in reservoir opossums. The proposed method could be used to control vector fleas among reservoir hosts and reduce risk for murine typhus in humans.