Variation in Angiostrongylus cantonensis infection in definitive and intermediate hosts in Hawaii, a global hotspot of rat lungworm disease

Abstract

Angiostrongylus cantonensis (rat lungworm) is a tropical and subtropical parasitic nematode, with infections in humans causing angiostrongyliasis (rat lungworm disease), characterized by eosinophilic meningitis. Hawaii has been identified as a global hotspot of infection, with recent reports of high infection rates in humans, as well as rat definitive and snail intermediate hosts. This study investigated variation in A. cantonensis infection, both prevalence and intensity, in wild populations of two species of rats (Rattus exulans and R. rattus) and one species of snail (Parmarion martensi). An overall infection prevalence of 86.2% was observed in P. martensi and 63.8% in rats, with R. exulans (77.4%) greater than R. rattus (47.6%). We found infections to vary with environmental and host-related factors. Body mass was a strong predictor of infection in all three species, with different patterns seen between sexes and species of rats. Infection prevalence and intensity for R. exulans were high in May 2018 and again in February 2019, but generally lower and more variable during the intervening months. Information on sources of variability of infection in wild host populations will be a crucial component in predicting the effectiveness of future disease surveillance or targeted management strategies.

Keywords: Angiostrongyliasis; Parmarion; Rattus; drivers; eosinophilic meningitis; nematode; parasite; slug; snail; transmission.

Fig. 1.

Model predictions for relationships among sex, mass and site for prevalence and intensity of Angiostrongylus cantonensis infection in Rattus rattus (A & C) and R. exulans (B & D) sampled at two sites in Hilo, Hawaii. Shaded areas represent ± 1 standard error for the estimate; where error intervals broadly overlap, there is little effect size.

Fig. 2.

Variation in Angiostrongylus cantonensis infection prevalence among sexes, sites and sampling periods for Rattus rattus (A) and R. exulans (B) sampled in Hilo, Hawaii. The solid line at the centre of each column depicts the proportion of rats infected in each subsample, irrespective of mass, and the shaded boxes represent the 95% binomial confidence intervals for the estimate. The number above each sample is the sample size (N). Periods A–D represent host sampling periods (A: May 2018; B: August–September 2018; C: November–December 2018; D: February 2019; with exact dates included in Supplementary Table S1.).

Fig. 3.

Model-predicted variation in Angiostrongylus cantonensis infection intensity for Rattus exulans among sampling periods. Shaded areas represent ± 1 standard error for the estimate; where error intervals broadly overlap, there is little effect size. Periods A–D represent host sampling periods (A: May 2018; B: August–September 2018; C: November–December 2018; D: February 2019; with exact dates included in Supplementary Table S1.).

Fig. 4.

Prevalence of Angiostrongylus cantonensis infection in Parmarion martensi snails sampled at two sites over four sampling periods in the vicinity of Hilo, Hawaii. Horizontal lines are the prevalence estimates by grouping, with grey bars indicating the 95% binomial confidence intervals, with values listed on the left y-axis. White circles represent the mean mass of snails in the sample, with vertical lines representing the 95% confidence intervals for the means and values listed on the right y-axis. Values above each boxplot indicated the sample size (N). Periods A–D represent host sampling periods (A: May 2018; B: August–September. 2018; C: November–December 2018; D: February 2019; with exact dates included in Supplementary Table S1.).

Fig. 5.

Model-predicted prevalence of Angiostrongylus cantonensis infection in Parmarion martensi snails sampled over four sampling periods at two sites in Hilo, Hawaii. The shaded area represents ± 1 standard error of the estimate.

Fig. 6.

Intensity of Angiostrongylus martensi infection in Parmarion martensi snails sampled over four sampling periods at two sites in the vicinity of Hilo, Hawaii, as predicted by the top model (mass + site + period + site × period). Shaded areas represent ± 1 standard error for the estimate; where error intervals broadly overlap, there is little effect size. Periods A–D represent host sampling periods (A: May 2018; B: August–September 2018; C: November–December 2018; D: February 2019; with exact dates included in Supplementary Table S1.).